The Getty Conservation Institute

C-ol-O-f- S c-i en ce i-n,----..--

of Museum

Ruth Johnston-Feller
Tools for Conservation
Color Science in
the Examination of
Museum Objects
Nondestructive Procedures
Ruth Johnston-Feller
The Getty Conservation Institute Los Angeles
Tevvy Ball: Project Editor
Elizabeth Maggio: Copy Editor
Anita Keys: Production Coordinator
Hespenheide Design: Designer
Printed in the United States of America
10 9 8 7 6 5 4 3 2 1
2001 The]. Paul Getty Trust
All rights reserved
The Getty Conservation Institute works internationally to advance conservation
in the visual arts. The Institute serves the conservation community through sci-
entific research, education and training, model field projects, and the dissemina-
tion of information. The Institute is a program of the J. Paul Getty Trust, an
international cultural and philanthropic institution devoted to the visual arts and
the humanities.
The Institute's Tools for Conservation series provides practical scientific proce-
dures and methodologies for the practice of conservation. The series is
specifically directed to conservation scient ists, conservators, and technical experts
in related fields. Previously published in the series are Infrared Spectroscopy in
Conservation Science (1999) by Michele R. Derrick, Dusan Stulik, and James M.
Landry, and Thin-Layer Chromatography for Binding Media Analysis (1996) by
Mary F. Striegel and Jo Hill.
"The Colors Live," from Hailstones and Halibut Bones, by Mary O'Neill and
Leonard Weisgard, Ill. Copyright 1961 by Mary LeDuc O'Neill. Used by
permission of Doubleday, a division of Random House, Inc.
Library of Congress Cataloging-in-Publication Data
Johnston-Feller, Ruth, 1923-2000.
Color science in the examination of museum objects : nondestructive
procedures I Ruth Johnston-Feller
p. cm. - (Tools for conservation)
Includes bibliographical references and index.
ISBN 0-89236-586-2
1. Art-Conservation and restoration. 2. Spectrophotometry.
3. Colorimetry. 4. Pigments. 5. Color in art. I. Title. II. Series.
N8560.J64 2001
702'.8'7-dc21 2001-031449
CIP
The colors live
Between black and white
In a land that we
Know best by sight.
But knowing best
Isn't everything,
For colors dance
And colors sing,
And colors laugh
And colors cry-
Turn off the light
And colors die.
-Mary O'Neill, "The Colors Live"
Contents
ix Foreword
xi Preface
xiii Acknowledgments
Introduction
Chapter 1 5 Spectrophotometry
6 Spectrophotometers
11 Reference Standards
12 Instrument Calibration and Measurement Reproducibility
13 Spectrophotometric Curves
Chapter 2 15 Colorimetry
15 Additive Color Mixture, Subtractive Colorant Mixture
18 Primary Colors
20 The CIE System
29 Examples
33 Co/or-Difference Equations
38 Tristimulus Filter Colorimeters
39 Additional Comments
40 Metamerism
47 Geometric Metamerism-a Special Case
48 Discussion
49 Other Color Notation Systems
49 Munsell System
53 Universal Color Names and Dictionary
56 OSA Uniform Color Scales
56 Other Systems
Chapter 3 58 Colorant Characteristics
Chapter 4 62 Colorant Mixture
62 Transparent Materials: The Beer-Bouguer Equation
62 Bouguer's (Lambert's) Law
64 Beer's Equation (the Beer-Bouguer Equation)
67 Absorbing and Scattering Materials: The Kubeika-Munk Equation
70 Qualitative Application of the Kubeika-Munk Relationship
78 Quantitative Application of the Kubeika-Munk Equation
78 Simple Estimation
86 More Complex Estimation
93 Applications of Kubeika-Munk Formulas
93 Study of Colorant Changes
102 Vehicle (Medium) or Substrate Change
106 Glazes
114 Opacity, Translucency, and Hiding Power
114 Definitions
116 Opacity Calculations
118 Scattering, Simple and Multiple
124 Charts for Determining Opacity
125 Hiding Power
128 Tinting Strength: Absorption and Scattering
145 Comments on Colorant Calculations and Identifications
145 Quantitative Limitations in the Use of the Kubeika-Munk Equation
147 Qualitative Applications of the Kubeika-Munk Relationship
149 Reflection Density
154 Special Scales and Methods Used in Industry
155 Applications in the Paper Industry
157 Other Single-Number Color Scales
Chapter 5 159 Color in Specular (Mirror-Type) Reflection
159 Color of Metals (Nondielectrics)
164 Bronzing
169 Pearlescence and Iridescence
172 History of Pearlescent and Iridescent Pigments
179 Other Flake Pigments (Metallic Flakes)
Chapter 6 190 Special Topics
190 Surface Reflection
190 Refractive Index Differences: The Cause of Surface Reflection
192 Matte Surfaces that Exhibit Geometric Metamerism
196 Surface Effects on Low-Chroma and High-Chroma Colors
203 Types of Gloss
205 Surface Changes after Exposure
205 Fluorescence
206 Color Measurement of Fluorescent Materials
207 Historical Use of Fluorescent Pigments
210 Development of Modern Fluorescent, High-Visibility Pigments
214 Uses of Fluorescent, High-Visibility Pigments by Artists
217 Fluorescent Whitening Agents (FWAs)
218 Weak Fluorescence of Resins
219 Microvoids and Vesiculated Beads
221 Extenders, Fillers, Inerts
Chapter 7 225 Reflectance Curves of Some Frequently Encountered Chromatic Pigments
225 Primary Colors
225 Blues
227 Reds
229 Yellows
231 Secondary Colors
231 Greens
233 Purples
233 Oranges
234 Pigment Interactions
237 Special Technique for Dark or High-Chroma Colors
Chapter 8 238 Measured-Data Analysis and Special Measurement Problems
Chapter 9 244 Instrumentation Overview: The Tasks Determine the Selection
246 Material (Sample) Characteristics
250 Other Instrument Features
251 Reports
Chapter 10 252 Suggested Protocol for Recording Spectral Examination Results
Chapter 11 257 Summary, Conclusions, and Recommendations
260 Appendix A: Curves and Data for Pigments Used as Illustrative Problems
294 Appendix B: K/S versus %R Table
296 Appendix C: Colour Index Name
298 Appendix D: Compilation of Spectral Reflectance Curves of Pigments:
Representative Example
328 Appendix E: Recommended Reading
332 References
351 Index
363 Illustration Credits
365 About the Author
Foreword
One of the principal objectives of the Getty Conservation Institute is to
advance the application of scientific principles and methods in the study
and care of museum collections. This book has been prepared to assist
conservators and color scientists in the measurement, description, and
investigation of the color and appearance of works of art. We are pleased
that, over the decade of her semiretirement, Ruth Johnston-Feller under-
took this project, which she was able to bring to completion in spite of
her steadily declining health. We deeply regret that she did not live to see
the present volume into print.
Fortunate is the individual who finds satisfaction in the work
that each day brings. The joy that conservators experience in their cho-
sen profession informs the myriad tasks that they perform in caring for
the irreplaceable materials in their charge. Ruth Johnston-Feller, a scien-
tist trained in chemistry, enjoyed the world of color with similar passion.
She also enjoyed the privilege of being present at the beginning of a new
and growing field-the field of color instrumentation and color science.
Over the years she taught numerous courses and generously shared her
experience and expertise, serving as a mentor for many colleagues and
students. In this monograph, she looked forward to the opportunity to
make accessible a body of knowledge gained over the course of a
lifetime.
During the years of preparation of this book, several authori-
tative texts on color science have appeared, including Anni Berger-
Shunn's Practical Color Measurement (1994) and a classic work that has
recently appeared in a third edition under the title Billmeyer and
Saltzman's Principles of Color Technology, ably revised and updated by
Roy S. Berns (2000). Each of these volumes, as well as the present mono-
graph, emphasizes somewhat different facets of color science application.
However, these distinguished authors, who worked together on a number
of projects, all emphasize the overriding importance of the fundamental
principles of color measurement and appearance.
Today technology rushes ahead at an ever-accelerating pace.
The cumbersome Leres spectrophotometer that was used to record many
of the spectrophotometric curves in this book, for example, has been
replaced by a variety of handheld devices that rapidly yield sets of
numbers representing both colorimetric and color-difference values.
Instead of unquestioningly accepting the numbers thus obtained, however,
x Foreword
these authors recommend that conservators make every effort to
understand the principles behind such data . Indeed, as this monograph
demonstrates, a solid grasp of the principles of colorimetry and spectro-
photometry provides a powerful tool for conservators and conservation
scientists to better understand a wide range of color-related problems,
such as the characterization and identification of pigments; changes in the
gloss, color, and translucency of paints; and the discoloration of binding
media that can occur with age. No other book on color science and mea-
surement, in fact, focuses as directly as the present volume on issues per-
tinent to the conservator. There are sections that will be useful for the
analysis and treatment of traditional materials as well as of materials used
in contemporary art. Other topics covered include the color measurement
of fluorescent materials, the effects of whitening agents, and the weak
fluorescence of binder resins. The book also treats subjects not commonly
found in the conservation literature, including the color of metals and of
pearlescent, iridescent, and metallic flake pigments.
We hope that this volume will assume a place among the
standard references in the conservation field. Ruth Johnston-Feller's life-
long enthusiasm for her work and her many insights into color science
constitute a distinguished legacy, one that the Getty Conservation
Institute is privileged to place on record.
Timothy P. Whalen
Director
The Getty Conservation Institute
Preface
Many fine books have been written about the science of color, about the
origin of color, about the techniques of its measurement and description,
about the interrelationship of color with other aspects of appearance,
and about sensation and perception. Some of these books are listed as
recommended reading in appendix E, and the reader is urged to consult
them. However, few books have been written that describe the applica-
tion of color science to the solution of practical problems. It is the pur-
pose of this book to present some examples illustrating how the basic
principles of color science can contribute to an understanding of the
color and appearance of materials. Emphasis is placed on nondestructive
analytical methods based on measurement and computation, and on how
the resultant information can be interpreted.
My experience in industrial applications necessarily forms the
basis for many of the discussions presented. I was employed by a number
of corporations with diverse interests in color science, and this experi-
ence provided me with a varied background of knowledge. These corpo-
rations included the Coatings and Resins Division of PPG Industries,
manufacturer of paints and plastics; the Macbeth Division of Kollmorgen
Corporation, manufacturer of color-measuring instruments and origina-
tor of computer software for color matching; and the Pigments Division
of Ciba-Geigy Corporation (now the Colors Division of Ciba Specialty
Chemicals), manufacturer of pigments. Many of the examples presented
in this volume had their origins in my experience with one or more of
these employers. I wish to express my thanks to these corporations for
their permission to use their equipment in my spare time, to work with
Robert Feller on a number of challenging conservation-science problems.
Because of this industrial background, many of my examples
as well as the materials I describe are taken from problems encountered
while performing day-to-day work. However, my introduction to
conservation-science problems provided an additional opportunity
to look at color science and its potential applications from a different
point of view. Beginning in 1975, I have worked on a consulting basis
in conservation science with Robert Feller at the Research Center on
the Materials of the Artist and Conservator, Carnegie Mellon Research
Institute, Carnegie Mellon University. Throughout my years of interest
in the problems of conservation, I have encountered many diverse appli-
cations for color science, some of which are described here.
xii Preface
In the pages that follow, emphasis is placed on the appropri-
ate use of spectrophotometry, which is the analytical tool for defining the
characteristics of color, and on the complementary use of colorimetry,
which is the numerical, three-dimensional specification of the visual
stimulus for the sensation of color. The examples presented are inter-
preted in terms that utilize information from both types of approaches,
emphasizing the complementary values of these two types of study.
The successful use of the methods described here depends on
the knowledge and skill of the persons using the instruments and inter-
preting the measurements. All of the important aspects of good analytical
science techniques are as basic to the application of color science as to
any other analytical procedure. At the very heart of all reliable analytical
work are proper instrument calibration; a thorough knowledge of instru-
ment limitations; multiple samplings and measurements to provide for
statistically reliable results; recognition of the effect of sample character-
istics on the method of analysis; and the need for reference information.
A plea is made here for the start of an ongoing program to build an atlas
of the spectral curves of colorants, both pigments and dyes. The need is
not only for colorants used in the past but also for those currently used;
such data would be of inestimable value for the future.
Because my experience has been most extensive in studying
pigmented systems, the reader will find that most of the examples pre-
sented are for those materials. However, many of the principles described
apply equally to dyed materials.
The list of references for this monograph is not as compre-
hensive as it could be. I wish to apologize to the many authors of valu-
able articles not specificall y mentioned. However, many excellent
supplementary articles can be found in the references cited throughout
and in the recommended reading in appendix E. Hopefull y, some older
and very valuable references to the basic work done many years ago will
be consulted by interested readers. We must never forget that we all
stand on the shoulders of the giants who came before us.
Ruth Johnston-Feller
Acknowledgments
This project was undertaken with the help and support of the Getty
Conservation Institute. Frank Preusser, former associate director of pro-
grams, arranged for the task to be undertaken; James R. Druzik gave
advice and encouragement throughout; and Eric Hansen made many
helpful comments.
The Center on the Materials of the Artist and Conservator at
the Carnegie Mellon Research Institute, headed by Paul Whitmore, spent
a large amount of time and effort on carrying out measurements and cal-
culations. The author appreciates the interest and help provided by Paul
Whitmore and his staff. Thanks are expressed to John Bogaard for help-
ing with references to the literature on the color of paper and on TAPPI
testing methods. Catherine Bailie is especially thanked for her laboratory
work, much of which is represented here; for her organization of the vast
collection of measurements, calculations, and spectral curves; and for
her liaison with the Photo and Graphic Services of Carnegie Mellon
University. Gary Thomas, head of Photo and Graphic Services, and Kelly
Young, who reproduced the spectrophotometric curves to a common
scale and produced the diagrams, played a most significant role in creat-
ing the illustrations so necessary in explaining the text. Thanks are also
given to Lynn Labun, librarian at the Carnegie Mellon Research Institute,
and to her staff for their assistance in locating many of the articles and
patents referenced in this monograph.
The help of Hugh Davidson of Davidson Colleagues in sup-
plying many of the computer programs, as well as in giving helpful sug-
gestions, is especially appreciated.
The valued criticisms, comments, and suggestions made by
Fred W. Billmeyer Jr. are acknowledged and greatly appreciated.
Many thanks are expressed to Max Saltzman for his interest,
help, and encouragement. And to Dennis Osmer, of Ciba Vision Corpo-
ration, the author wishes to express her thanks for his continuing interest
and many helpful discussions.
The author would also like to acknowledge the help of Leon
Greenstein in preparing the section on pearlescence and iridescence.
The author is grateful to the many copyright holders for their
kind permission to utilize and reproduce graphs and drawings that were
considered especially useful in illustrating particular ideas in the text.
xiv Acknowledgments
To Sandra Melzer, secretary at the Carnegie Mellon Research
Institute, the author expresses her gratitude for her continued help in
typing, correcting, and organizing the many drafts, all of which was done
with great skill, speed, and patience.
Most of all, the author wishes to acknowledge the continuing
interest and help of her husband, Robert L. Feller, with whom she has
worked on, discussed, and studied much of the material presented. His
aid in the organization and logistics, and his continuing support, have
been invaluable.
In the end, it is the editorial and design work of Getty Trust
Publications that, in cooperation with the author, has produced the
present volume.
Introduction
Care of the art and artifacts produced by civilizations past and present is
a responsibility owed to posterity-to our children, to their children, to
their children's children. To carry out this trust, every effort must be
made to preserve cultural materials with the best possible knowledge and
skill. To a great extent, their proper care depends on identification of the
materials used in their creation. Of the many analytical techniques that
can be employed for colorant identification, those that do not require
the taking of samples should be used first. One such noninvasive tech-
nique is spectrophotometry, which measures the amount of light reflected
or transmitted by a material at individual wavelengths of the spectrum. A
graph of the relative percent of electromagnetic power reflected or trans-
mitted at each wavelength is called a spectrophotometric curve or simply
a spectral curve. The shapes of such curves are often unique for various
colorants-pigments or dyes-and thus can be helpful in identifying the
colorants present in the object. The discussion in these pages is limited to
the spectral curves measured in the visible spectral region. These are the
curves responsible for the color we see. Although positive identification
may not always be possible with spectral curves, information concerning
what is not present can also be of considerable value.
We do not see spectral curves, however. Our eyes are not ana-
lytical instruments-they are integrative. We hear analytically, that is, we
hear the single frequencies (or wavelengths) of sounds in chords and as
harmonies. On the other hand, we do not see individual wavelengths (or
frequencies) of light, only their summation. (Frequency of radiant power
is the inverse of the wavelength, so the two terms are inexorably related.)
This integration, or summation, of radiant power over all wavelengths of
visible radiation results in a single color response: blue or green, light or
dark, brilliant or dull.
There are many factors, in addition to the pigments and dyes
present, that affect our perception of an object's color. It is first of all
affected by the light source- a red object looks black in green light, for
example, because all of the incident green light is absorbed. Moreover,
each observer may see colors somewhat differently because humans differ
in their responses to wavelength, and people with the condition popu-
larly referred to as color blindness may vary radically in their responses
from those with so-called normal color vision. The science of colorimetry
includes the wavelength intensities of a number of standard illuminants-
2 Introduction
selected by the Commission Internationale de l'Eclairage, or the
International Commission on Illumination (CIE)-and the wavelength
sensitivities of two average observers. One standard observer represents
distance viewing, and one represents arm's-length viewing. By suitable
measurement of the reflectance of an object (or transmittance of a trans-
parent material) followed by calculations incorporating the data for one
of the CIE standard illuminants and a standard observer, three figures
can be obtained that can be used to describe any color. This science of
colorimetry, under the aegis of the CIE, is accepted worldwide. The CIE
system provides a standard procedure for describing a color stimulus in
terms of defined illuminants and a defined standard observer.
The sensation of color is complex. Perception of a scene
involves many variables. Considered analytically, color sensation incor-
porates not just the factors of object coloration but also surface gloss
or matteness, illuminant directionality or diffuseness (which affects our
evaluation of surface characteristics), and size and shape. Perceptually,
viewing a complex scene further depends on other factors difficult to
characterize, such as the contrast between neighboring colors and the
adaptation of the eye to the total surroundings. These factors make it
difficult to match any single color in a complex composition on the basis
of our visual impression. It is physical measurement, made under defined
geometrical conditions, that can provide an unambiguous description of
an object's reflectance or transmittance. This objective measurement, by
its revealing wavelength characteristics, can aid in the identification of
the colorants involved.
It is therefore necessary that the study of colored objects
utilize both the analytical approach of spectral analysis and the colori-
metric data obtainable by use of the methods described by the CIE. Both
approaches are used in the descriptions and applications of color science
discussed in this monograph.
Spectrophotometry in the visible wavelengths and the science
of colorimetry can be useful in many important ways in objective analysis
and research. These include monitoring changes in materials following
exposure to light or other deleterious conditions; measuring the degree
of yellowing of a vehicle, varnish, or substrate; documenting changes in
gloss and the resulting effect on color; determining the degree and rate
of colorant fading or darkening; and measuring the colors of metals and
their change upon exposure. One of the chief purposes of the remarks that
follow is to describe how these particular problems can be approached.
Fascinating new pigmentary materials, such as microvoids
and vesiculated beads, came into use following World War II. Many of
these new materials-fluorescent pigments, metallic pigments, and irides-
cent pigments-present special spectral measurement problems and are
discussed in chapters 5 and 6.
It is not the purpose of this monograph to provide another
textbook on color and spectrophotometry. In the initial sections a few
basic principles are reviewed so that the reader will be reminded of the
fundamentals used in the various applications described in subsequent
chapters. The primary emphasis, however, is on the description of
Introduction 3
methods of approach for studying a variety of problems and materials
and on understanding the principles involved so that, even without elab-
orate equipment, valuable information can be obtained. It is hoped that
the guidelines presented will help prevent the occasional misinterpreta-
tion of the results of spectrophotometric and colorimetric measurements
and the drawing of erroneous conclusions from them. As with any ana-
lytical technique, reference materials of defined composition are the first
essential; so it is with using spectrophotometric curves. In the discussions
that follow, it should become evident that an atlas of curves of known
colorants is a necessity.
It is assumed that the reader has some technical background
but is not a specialist in color science and its applications. A certain
knowledge of basic mathematics is necessary. This monograph has been
prepared primarily for museum laboratory personnel. Nonetheless, it is
hoped that others also may find some of the ideas, discussions, and
applications helpful.
Chapter 1
Spectrophotometry
The optical instrument used to measure how materials reflect or trans-
mit light is called a spectrophotometer. Such instruments measure the
relative amount of electromagnetic power reflected or transmitted at
individual wavelengths of the spectrum. When the amount of power
reflected or transmitted by a material is scanned across a particular
region of the spectrum, a curve may be plotted showing reflection or
transmission as a function of wavelength. Such a curve is called a spec-
tral curve or a spectrophotometric curve. The shape of this measured
curve is often specific, providing a type of fingerprint characteristic of
the chemical nature of the material. This fingerprinting and objective
characterization of colored materials in the visible spectral region is one
of the subjects of this monograph.
The unit of wavelength used in the visible region is the
nanometer, which is equal to 10-
9
meters and is abbreviated nm. (Older
literature used the term millimicron, abbreviated m, for the same size.)
In plotting graphs of reflectance or transmittance versus wavelength, the
x (horizontal) axis is the wavelength, often indicated by the symbol A,,
and the y (vertical) axis is conventionally the reflectance or transmit-
tance, generally expressed as a percentage on a scale of 0 to 100. The
inverse of the wavelength, lit..,, is the frequency, which is often used in
the infrared region. The CIE currently recommends using the term
reflectance factor for the ratio of the light reflected by a sample relative
to that reflected by a reference standard. This is described in Standard E
284 of the American Society for Testing and Materials (ASTM). For
convenience, however, the older term percent reflectance (or simply
reflectance) is used throughout this monograph, even though, in modern
CIE terms, reflectance factor is being measured. In these pages there is an
additional word modifying reflectance: either diffuse, meaning "specular
reflectance excluded," or total, meaning "specular reflectance included."
To describe color in CIE terms, diffuse reflectance must be used. To ana-
lyze the intensity of the light beam, the total reflectance has to be consid-
ered. Illustrations of spectrophotometric reflectance curves for various
pigments in glossy paints are presented in appendix A and in chapter 7.
They are used frequently as examples throughout this book.
6 Chapter 1
Spectrophotometers
All spectrophotometers consist of three essential parts: ( 1) a light source,
(2) a monochromator to select the individual wavelength, and (3) a
photodetector. The light source for visible wavelengths is generally either
a tungsten lamp, a tungsten-halogen lamp, or a xenon flash-tube lamp.
The monochromator is either a grating or a prism, suitably equipped
with a slit mechanism to isolate a narrow band of wavelengths. The
detector is a device that is sensitive to radiation in the desired visible
spectral region, such as a silicon photodiode or a photomultiplier tube.
A photodiode array is often used today. Additional essential parts of a
spectrophotometer are arrangements for mounting and illuminating the
sample, a reference standard, and a geometric arrangement for selecting
or collecting the light that falls on the detector. The measurement made
with a spectrophotometer is the ratio of the light reflected, or transmit-
ted, by the sample to that reflected, or transmitted, by a reference stan-
dard. This ratio can be described as I/I0, where I is the light intensity
from the sample and I 0 is the intensity from the reference material. It
may be recorded by a plotter as a continuous (analog) curve or printed in
digital (numerical) form. The advantage of the digital output is that any
corrections to the data can be made to correct for instrument calibration
deviations or to calculate the measured ratio to a different reference stan-
dard. Such corrections are desirable to achieve maximum precision or
accuracy. Color-measuring instruments available today are equipped with
computers to make such corrections.
Most modern instruments do not measure continuous (ana-
log) curves, such as those illustrated in appendix A and chapter 7. For
example, color-measuring instruments produced for industrial color
control measure at 10 nm intervals, or even just 20 nm intervals. Such
digital instruments are now popularly called spectrophotometers, but
they are only "abridged" spectrophotometers, as they were formerly
described. Data from such instruments cannot provide a continuous
analog curve. Their advantage for use in industry is the speed of mea-
surement. With the use of diode-array detectors, 33 wavelength measure-
ments can be made in a few seconds and entered into a computer for
calculations, such as color differences or colorant formulations, which
are then printed out. In addition, corrections for instrument deviations
in calibration can be calculated before the measurements are utilized.
Often, however, the operator does not even see the spectrophotometric
data on which the derived calculations were based and thus has no idea
of the value of the spectral information being missed. Berger-Schunn
(1994) expressed the present author's views: "While the reflectance
curves in the early days of color measurement had to be drawn by hand
or, when automatically measured, at least had to be looked at for further
use, with modern color measurement systems the curves are measured
but often neither printed nor recorded. The user today therefore may
miss a large amount of information."
Modern instruments that measure the reflectance or transmit-
tance in digital terms but at 2 nm or less wavelength intervals can prop-
Spectrophotometry 7
erly be called spectrophotometers. Graphs made of the percent reflec-
tance or transmittance of such digital data are very similar in curve shape
to the older analog continuous curves measured on the best spectro-
photometers that were used to record the spectral curve characteristics
of many colorants in the post-World War II period until the late 1970s.
Many of these valuable curves still exist and should be accumulated in
an atlas for future reference.
The characteristics of the sample to be measured will deter-
mine the selection of a spectrophotometer's essential components and
their most suitable arrangement. The cost and availability of the equip-
ment must also be considered. Throughout the discussions that follow,
reference is made to the instrument requirements for various types of
samples and to the measurement tasks. Chapter 9 is devoted to a dis-
cussion of defining the problem to be solved and selecting the proper
instrument to meet its needs. ASTM Standard E 805 presents a concise
summary of the features of spectrophotometers, as well as of colorime-
ters, that need to be considered when making measurements of the color
or color difference of materials.
Two basic differences in instrument design should be noted
here at the start. The first involves the location of the light source and
the monochromator relative to the location of the sample. In some
spectrophotometers, light from the source is first passed through the
monochromator to isolate narrow wavelength regions. The resulting
monochromatic light then illuminates the sample in the sequence:
lamp monochromator sample detector. However, if the sample
exhibits any fluorescence, the measurements made with this arrangement
cannot accurately represent its visual appearance. This is because any
emitted fluorescence is registered by the detector at the excitation wave-
length rather than at the emission wavelength. (See discussion of fluores-
cence in chap. 6.)
An arrangement whereby the locations of the sample and the
monochromator are interchanged is now more commonly used in color-
measuring instruments in the sequence: lamp sample monochroma-
tor detector. This arrangement provides measurements agreeing with
what a human observer sees. With this interchange of the sample and
monochromator positions, the light reflected or transmitted by the sample
passes through the monochromator after being reflected or transmitted
by the sample. The sample is illuminated by the polychromatic light from
the source. The spectral distribution of this polychromatic source is criti-
cal when measuring fluorescent samples; the power distribution of the
excitation wavelengths influences the amount of the fluorescence emitted.
When measuring nonfluorescent samples, either arrangement of sample
and monochromator is satisfactory.
The second important aspect of spectrophotometer design
relates to the geometric arrangement for illuminating the sample and
for collecting the reflected light. Reflection at the air-sample interface
always occurs because of a change in the refractive index. (Objects
would be nearly invisible if this were not true. ) On highly polished or
very glossy materials, the surface reflection is mirrorlike, that is, light
8 Chapter 1
will be reflected at an angle equal and opposite to the incident angle.
This type of mirrorlike reflection is called specular or Fresnel reflection.
When surfaces are matte (nonglossy), all of the surface reflection is dif-
fuse-that is, the light is scattered in all directions. Between these two
extremes there are all degrees of glossiness or matteness. In dielectric
materials (nonmetals), the surface reflection is simply the character of the
illuminant. Color is produced from the internal absorption and scattering
of the incident light inside the material. Therefore, to describe the color
as observed, it is desirable to exclude the specular reflection on glossy
samples. This is what human observers do; they ignore the very bright
highlights. On matte samples, the diffusely reflected surface light with
the spectral character of the illuminant is mixed with the internally
reflected light so that the observer sees the object color as desaturated or
less pure. Metals (nondielectrics), on the other hand, reflect only at the
surface. Light does not significantly penetrate into a metal. The color of
copper and gold, for example, comes from the surface reflection alone.
(See discussion of metal color in chap. 5.)
For collecting reflected light, three basic types of optical
design are used in spectrophotometers: ( 1) bidirectional geometry, gener-
ally 0 (perpendicular) illumination, 45 viewing (written as 0/45, or
the reverse, 45/0); (2) integrating sphere geometry with either 0 illumi-
nation and diffuse viewing (0/d) or near-normal incidence, such as 8,
and diffuse viewing (8/d); and (3) variable angle geometry in which the
incident and viewing angles may be changed.
The first type, bidirectional geometry, measures only diffusely
reflected light; the specularly reflected light is reflected back into the
instrument. Such an instrument is unsuitable for measuring the color of
metals and, often, is not designed to measure the transmission of light
by transparent materials. Its only purpose is to measure the diffusely
reflected light-the color-of dielectric materials.
The second type, integrating sphere geometry, is more versa-
tile. However, if the angle of illumination is normal (perpendicular, or
0) to the surface of the sample, the specular surface reflection is also
excluded and reflected back into the instrument . This type of integrating
sphere with its rejection of the specular reflection is inappropriate for
measuring the color of metals. In the most commonly used type of inte-
grating sphere geometry, however, the angle of illumination is about 8
off the normal, or perpendicular (0), so that light reflected specularly
at -8 (that is, at 8 but on the opposite side of the perpendicular, as
indicated by the minus sign) strikes the integrating sphere wall and is
included in the measurement. Such instruments are appropriate for mea-
suring the color of metals and also for making transmission measure-
ments. They are generally equipped to exclude the specular reflection,
if desired. For example, at the angle equal and opposite to the incident
angle of illumination, a piece of the sphere wall can be removed, and a
black port (or light trap) inserted in place of the white port. With the
black port in place, the specularly reflected light will be absorbed, and
only the diffusely reflected light is recorded (see fig. 6.9b).
Spectrophotometry 9
Variable angle geometry, the third type of optical design, uti-
lizes variable angles of illumination and view. Such an arrangement is
extremely valuable for measuring materials that change color when the
viewing angle or the illumination angle is changed. Some materials that
exhibit such color changes are those containing metallic flake pigments
or iridescent pigments (see chap. 5). Several recent commercial instru-
ments use this design. Two basic types of instruments that measure color
as a function of the angles of illumination and view can be distinguished.
In one optical arrangement, the reflection is measured as a function of
the angles of illumination and view, keeping the wavelength of illumina-
tion constant. Such an instrument, a goniophotometer, can be called a
spectrogoniophotometer if provision is made for changing the wave-
lengths of the incident light across the spectrum (FSCT 1995; ASTM
Standard E 284 ). Reducing the measurements made by such instruments
to color terms requires the use of many wavelengths of incident light
across the spectrum. A computer, which is an essential part of such
instruments, can carry out these calculations for each wavelength of illu-
mination and angle of view. Any variable angle measurement can be
invaluable in characterizing the nature of surfaces as well as of internal
structures. However, if one is interested primarily in the differences in
color as a function of the viewing angle, an instrument that measures the
spectrophotometric curves at specified angles of illumination and view is
more convenient. Such an instrument is called a goniospectrophotometer
(FSCT 1995; ASTM Standard E 284).
The instrument referred to throughout this monograph is a
goniospectrophotometer, specifically the Trilac manufactured by Leres in
France. This is the instrument used for the variable angle measurements
described in chapters 5 and 6. The Trilac also has the more common
mode of optical arrangement utilizing an integrating sphere for measur-
ing diffuse and total reflectance in the conventional manner. It is no
longer manufactured.
In its goniospectrophotometer mode (hereafter referred to as
the "gonio mode"), the instrument does not have high resolution; the
incident beam is nearly collimated (parallel beam), with about 4 aper-
ture angle. In contrast to a number of more modern gonio instruments,
however, the viewing angle is relatively wide, with an aperture angle of
about 15. This angle of view is close to visually subtended angles, with
the result that the measured color agrees moderately well with visual
observations. In all current commercial gonio instruments, including the
Trilac, the angles of illumination and view lie in the same plane, as illus-
trated in figure 1.1. With gonio measurements, it is easiest to interpret
the results if one angle, either that of illumination or view, is kept con-
stant as a reference point. More complicated gonio measurements usually
result in large numbers of measurements and are challenging to interpret.
The notation used in this monograph to describe the angles
of illumination and view is the following: The angle of illumination is
given first, designated as negative if on the side of the perpendicular
opposite to the angle of view. This is followed by a slash, and then by
10
Figure 1.1
The two planes as used in the Trilac
goniospectrophotometer. The shaded
plane represents the incident and viewing
plane, perpendicular and at right angles to
the sample plane (in white). The viewing
angle (v) is positive. The incident angle (i)
is negative when on the opposite side of
the normal (0) line, or positive when on
the same side as the viewing angle. On
the instrument, the viewing-angle dial is
labeled a, and the dial representing the
total subtended angle, i + v, is labeled
Chapter 1
'
'
'
/
'
/
the angle of view, which is always positive. One further convention used
here is to give the aspecular angle (that is, the angular degrees of viewing
off the specular angle) in parentheses. Thus a t ypical notation to describe
the geometric conditions used in a measurement is: -30/10 (20) . This
describes an angle of illumination of -30 and an angle of view of 10,
which is 20 off the specular angle of 30.
For all of the measurements so described, the reference white
standard was pressed barium sulfate (BaS04 ) . This was also used for the
gonio measurements. When making variable angle measurements, only
perfectly diffuse samples measure the same reflectance at any angle of
illumination and view. Measurements on any other type of sample are
not absolute and so are designated as relative. Thus, for all of the curves
illustrated in chapter 5 (under " Bronzing," " Pearlescence and
Iridescence," and "Other Flake Pigments" ), the ordinate is designated as
relative reflectance. On nondiffuse samples, as the measurements at the
viewing angle approach the specular angle, the measured reflectance rises
sharply and can surpass 100% relative to the diffusely reflecting white
standard. To make measurements at the specular angle as described for
the pearlescent and interference pigments, a neutral-density filter is
required in the sample beam of the instrument to decrease the specular
reflectance to 10% or 20% of the true value. In this way the response
will stay on the instrument scale. It should be noted also that the CIE
notations calculated for such relative reflectance gonio measurements,
when converted into color-difference data, do not have the same meaning
as true absolute reflectance data measured on conventional nondirec-
tional samples with conventional color-measuring instruments.
Corrections can be made for the absorption of the neutral-density filters
used, however. Without such corrections, and assuming that the neutral-
density filter transmission curve is very close to neutral, the derived CIE
chromaticity coordinates-x and y-provide useful information for com-
parison with simi lar samples measured in similar fashion . In summary,
Spectrophotometry
such relative gonio measurements are not absolute in the normal sense,
as are measurements made with conventional color-measuring instru-
ments intended to measure diffuse or total reflectance.
Note that transfer standards (defined below) for the diffus-
ing reference white, such as white glass, should not be used unless they
have been calibrated for the particular angular conditions used or fully
described for comparison purposes.
Reference Standards
11
Regardless of the type of instrument used to measure transmittance or
reflectance, all measure the ratio of the light flux transmitted or reflected
by the sample relative to the flux transmitted or reflected by a reference
standard. In measuring transmittance, the reference standard may be air
or, in the case of solutions, an identical cell containing the solvent used
to dissolve the sample material. Correspondingly, when measuring
reflectance, a reference standard that reflects 100% at all angles over the
entire spectral range being studied is desirable. This material would be a
perfect reflecting diffuser. No such material has been found to exist,
although many materials come close to this ideal. The CIE has recom-
mended it as the official white reference to use for measurements made
to describe color. Since no such material exists, a transfer standard
(Johnston 1971a) is used. This is a white material that has been cali-
brated in absolute terms, compared to the perfect diffuser, by a national
standardizing laboratory, such as the National Institute of Standards and
Technology (NIST), formerly the National Bureau of Standards. Such cal-
ibrated physical standards are generally supplied with an instrument
when it is purchased, but they may also be purchased separately. They
are frequently made from white opal glass or other white materials and
are in general relatively permanent. Other standards may be used as
working standards for everyday use. Diffuse white materials, such as
pressed BaS04 and pressed poly(tetrafluoroethylene) (Halon G-50), have
been recommended. These "working standards" may be calibrated
against the transfer standard, and the results stored in the computer for
corrections, if desired. (See ASTM Standard E 259 .) A stack of Millipore
filters with a pore size of 0.025 m (for example, VSWP 293-25) can also
be recommended, particularly for use where pressed powders are inap-
propriate. These materials owe their whiteness to entrapped air spaces.
Many operators simply use their calibrated transfer standard as a work-
ing standard.
In considering what to use for a working standard, there is
one major characteristic of a working standard that, in the author's
experience, is paramount: its surface characteristics. Although relative
whiteness is important when making the most precise measurements, the
diffuseness-or lack of gloss-of the working standard can be even more
important. This is especially true when making measurements on materi-
als that exhibit color in the specular or mirrorlike reflectance, such as
bronzed paints or inks, or interference pigments (see chap. 5).
12 Chapter 1
If one wishes to make very precise measurements, all care must
be taken to use the best white reference available. However, one may ques-
tion how significant this is for everyday practical work of curve-shape
identification. All of the above-mentioned powder transfer standards, that
is, all transfer standards except the ceramic materials, reflect about 97%
or more on an absolute scale (relative to the perfect diffuser ). Thus, if a
pressed BaS04 standard has an absolute reflectance of 98%, the maximum
error is 2 % of the measured reflectances. At the 40% reflectance level,
instead of measuring 40.0%, one will measure 40.8 %. Does it really mat-
ter? No. Since the whole measurement scale is displaced, this is of little
consequence. When spectrophotometry is used to analyze the relative
shapes of curves, the absorption ratios at various wavelengths, which cre-
ate the characteristic fingerprints, are the important aspect.
Finally, it must be pointed out that there are, in reality, no
"absolute" measurements in spectrophotometry or in colorimetry. All
measurements depend ultimately on both the geometric and spectral
design of the instrument. For example, one will not obtain exactly the
same measurements on sphere instruments as on bidirectional (0145 or
4510) instruments (Raggi and Barbiroli 1993 ). Reference curves used for
identification need to be made on instruments on which similar curves can
be measured. In the pages to follow, continuous curves recorded with the
General Electric Recording Spectrophotometer or with the Trilac are pre-
sented. The white reference was a diffuse pressed white powder, BaS0
4
.
Except for the variable angle measurements, all of the reflectance curves
were measured using one of these two spectrophotometers, both equipped
with integrating spheres. Agreement between these two instruments was
excellent on all reflectance measurements. Today other instruments will be
used. They should be identified and described.
Instrument Calibration and Measurement Reproducibility
When it is desired to make color and color-difference measurements
with the reproducibi lity over time that approaches the sensitivity of the
human observer to see small color differences, the measurements must be
made with the utmost care. Instruments must be accurately calibrated.
The samples being measured must be handled carefully. Repetitive mea-
surements need to be made. Additionall y, in using any particular instru-
ment, the precision attainable should be known in order to interpret the
measured results. Regardless of the color-measuring instrument used, a
procedure for regularly calibrating it should be followed. Hemmendinger
(1983) gives an extensive review of spectrophotometer calibration proce-
dures. Basically, measurements of selected stable reference standards are
necessary for determining photometric and wavelength scale accuracies.
Also, an understanding of the means for adjustment and correction, if
necessary, is required.
From repetitive measurements of the stable reference stan-
dards, the reproducibility with which measurements can be made can be
determined. The process of operating color-measuring instruments reli-
Spectrophotometry 13
ably has been described in ASTM Standards E 1164, E 1331, E 1345,
E 1348, and E 1349. Most of the modern color-measuring instruments are
interfaced to a computer so that corrections are automatically applied to
the measured data. It is important that the results on the reference stan-
dards be studied and interpreted over periods of time if measurements
are going to be made over such periods of time. Obviously, studying fad-
ing is one such application for which the reliability of the measurements
must be known. The latest instruments are very reliable, but people are
not always so. If I use an instrument, I want to know how well I can
operate it. How significant are my data?
Periodically in these pages sample characteristics and their
effect on color measurement are discussed. But it is not simply the nature
of the sample that is important; the condition of the sample and of the
reference white standard are vital. Too often fingerprints appear and
scratches occur. Dirt accumulates. Such happenings are most likely to
occur when measurements are made over an extended period of time.
They can be responsible for unwanted errors. Every student of statistics
learns the importance of making multiple measurements. Measuring
color is no different. Multiple measurements are important. Treating
them statistically is discussed in chapter 8.
Spectrophotometric Curves
The value of the spectrophotometric curves presented throughout these
pages lies with the skill of the person interpreting the measurements.
Ultimately, the most important, and generally the most expensive, part
of the successful application of spectrophotometry as a tool in material
analysis is the person behind the instrument. With practice, operators
can learn to read spectral curves like another language; they may not
necessarily be able to interpret every word or nuance, but they can
understand the implications and limitations.
Spectrophotometric curves in the visi ble spectrum are gener-
ally plotted as percent reflectance or transmittance on the vertical axis
(y) versus wavelength on the horizontal axis (x) at a scale of 400-700
nm. The color of the wavelength regions is indicated at the top of the
spectrophotometric curves in this monograph, from violet at the left
through the spectrum to red at the right. If a sample reflects more blue
wavelengths than those in other regions, one soon comes to realize that
the material will look blue to an observer (see figs. A.8 and A.10). If the
sample reflects both blue and red (both ends of the visible spectrum), it
looks purple (see figs. A.25 and A.27). Yellows reflect green, yellow,
orange, and red, which visually mix additively to make yellow (see
chap. 2, "Additive Color Mixture"). Neutrals and near neutrals, such as
whites and grays, reflect all wavelengths, approaching a flat refl ectance
curve. The closer a curve comes to this fl at shape, the more it resembles
a neutral color, and the lower is its saturation or Chroma, which is the
Munsell designation that corresponds closely to saturation. (See chap. 2,
"Other Color Notation Systems.") An illustration of the effect of adding
14 Chapter 1
black to any chromatic color is shown in figures A.9, A.11, A.13, A.15,
A.17, A.19, A.21, A.23, and A.26. Conversely, the more contrast there
is between the minimum and maximum reflectance, the higher is the
saturation (Chroma). Another indicator of saturation is the steepness of
the curve Setween lowest and highest reflectance: the steeper the curve,
the higher the saturation will be. This is particularly evident in yellows,
oranges, and reds, where the steepness of the curves is most obvious.
Thus, the three dimensions of color-hue, saturation, and
lightness-can be inferred roughly from an examination of the spectral
reflectance curve measured on a specimen. Hue can be visualized approx-
imately from the reflectance maximum; saturation (or its opposite, gray-
ness) from the contrast between the reflectance minimum and maximum;
and lightness from the amount of the reflectance.
For analytical identification purposes, however, it is not the
nature of the spectral reflection but the character of the absorption (low
reflection region ) that is definitive. It is the region of low reflection where
the chemical structure of the colorant has its effect on the curve shape. So
it is in this low reflection region where we look for a fingerprint. To see
the characteristic absorption of pigments, they have to be diluted with
white pigment or applied as incomplete hiding over a reflecting substrate;
to see the absorption of textile dyes, they have to be diluted with solvent
(when in solution) or applied thinly on a reflective substrate such as a
light textile fabric. Throughout this monograph pigments are used as
examples and are mixed with white. The author is most familiar with pig-
ments as colorants. An examination of the curves in appendix A and in
chapter 7 illustrates how some of the pigments have distinctive absorption
curve shapes in mixture with white, whereas others have very little dis-
tinctive shape or clue to identify them. Many yellows, for example, are
not easily distinguished from one another, although there are subtle differ-
ences that can provide some help, such as the wavelength at which the
reflectance increases (or the absorption decreases).
Many topics are discussed in the pages that follow. Most
concern the use of spectral curves as aids in the identification of colored
materials. Of great help, however, is simply looking at the material! Our
eyes are very sensitive instruments, and they are basic to an intelligent
study of colorant composition. On the wall of every color-measurement
laboratory there should be a sign that says: LOOK: THINK. (This is
opposite in order to the admonition given by Billmeyer and Saltzman
[1981] in their Principles of Color Technology. ) Look at the samples.
Look at the spectral curves. Then think about what they tell you.
In the end, every nondestructive tool available for chemical
analysis should be used before taking samples of valuable materials. Our
visual impressions can be characterized with considerable objectivity by
calculating colorimetric descriptions from spectral curves using the sci-
ence of colorimetry, which is the subject of chapter 2.
Chapter 2
Colorimetry
Wyszecki and Stiles (1982) define colorimetry as the "branch of color sci-
ence concerned with specifying numerically the color of a physically
defined visual stimulus." Thus a system of colorimetry is basically a " lan-
guage" with which an observer may describe a color unambiguously and
uniquely to distinguish it from all others. Although the basic techniques
for determining colorimetric data are described in this chapter, the reader
is also referred for additional details to the recognized textbooks on
color science listed in appendix E.
Regardless of their basic type, all systems of colorimetry are
required to be three-dimensional because of the nature of human vision.
This means that the specification of three independent variables is
required to describe colors uniquely. A perceptual system of colorimetry
is based on visual appearance, with the three dimensions of hue, satura-
tion (distance from neutral), and lightness; a psychophysical system is
based on additive color mixture of three primary colors; and a physical
system is based on subtractive colorant mixture of three primary col-
orants. Most of the commonly used systems are based on one of the first
two types: perceptual or psychophysical. The last type, a physical system,
is exemplified by many commercial color atlases based on a limited num-
ber of colorants mixed systematically.
Additive Color Mixture, Subtractive Colorant Mixture
In the preceding paragraph, two common but confusing concepts that
require clarification were introduced: (1) color and colorant mixture, and
(2) primary colors and primary colorants. As children, we learned that
the three primary colors are red, blue, and yellow. This appeared to be
confirmed when, using little pans of watercolor, we mixed yellow and blue
and made green. However, we were not mixing color but rather colorants,
that is, pigments or dyes. This type of colorant mixture is called subtrac-
tive colorant mixture because the individual colorants absorb-that is,
subtract-portions of the incident light in certain spectral regions, leaving
only the light from the nonabsorbed spectral regions to be reflected or
transmitted, and observed. Thus, in the case of the mixture of yellow and
blue watercolors, the yellow absorbs the violet and blue regions of the
spectrum, and the blue absorbs the orange and red regions. As a result,
16
Figure 2.1
Spectrophotometric curves calculated for
equal mixtures of yellow, blue, and white
colorants mi xed subtractively (a physical
mixture) and additively (a visual mixture) .
The subtractive mixture (curve 1) is a
green; the additive mixture (curve 2) is a
near-neutral color.
Chapter 2
only green is left to be reflected. Subtractive mixture is of major concern
when identifying colorants and is discussed in chapter 4.
In contrast to the term colorant, the term color involves the
human response to light reaching the eye. Each of three color-vision
receptors in our eyes adds light of all the visible wavelengths into a
single response, hence the term additive color mixture. Thus vision is an
integrative process, and not an analytical one, in terms of the individual
wavelengths of light reaching the eye. (Conversely, as mentioned earlier,
hearing is an analytical process because it recognizes individual frequen-
cies of sound reaching the ear. )
Returning to the watercolor paints, if we painted very tiny
squares of yellow and blue, and looked at this array at such a distance
that we could not distinguish the individual squares, we would see the
additive mixture of the light reaching our eyes from the blue squares and
from the yellow squares. The color observed can be calculated by averag-
ing the spectral reflectance curves for the yellow and the blue since each
occupies half of the area. The spectral curve for the subtractive colorant
mixture of yellow iron oxide, phthalocyanine blue (hereafter abbreviated
"phthalo blue" ), and white (Ti02) is shown as curve 1 in figure 2.1, and
'El
g
"'
u
c
2
u
"'
"" 1:

c
"'
"' "-
- V-11-c--- B --.r-- G y 0 ----R----
90
80
70
60
50
40
I
30
I
20
10
400
25% yellow iron oxide, 75% Ti0
2
0.5% phthalo blue, 99.5% Ti02
equal mixtures
2
..
,
-----...
--------
,
I
,
--subtractive} 12.5% yellow iron oxide
0.25% phthalo blue
- - - - additive 87.25% Ti02
450 500 550
Wavelength (nm)
600 650
.. -
700
Figure 2.2
Spectral curves calculated for equal mix-
tures of an orange-red (2 % red oxide and
98% Ti0
2
) with a blue (0.5% phthalocya-
nine blue and 99.5% Ti0
2
). The subtrac-
tive mixture (curve 1) is a near neutral;
the additive mixture is a purple (from
Johnston 1973. Copyright 1973
John Wiley & Sons, Inc. Reprinted by
permission).
Colorimetry 17
the additive color mixture of the same materials applied as little squares
is shown as curve 2. Examination of curve 2 indicates that the resultant
color is a nearly neutral gray and is lighter than the subtractive green
mixture. Spectral curves for the individual colorants-phthalo blue
and white, and yellow iron oxide and white-are illustrated in figures
A.10 and A.18, respectively.
An additive mixture does not necessarily result in colors that
are nearly gray. The curves illustrated in figure 2.2 for the subtractive
and additive mixtures of red oxide and white, and phthalo blue and
white illustrate this in an interesting way. Red oxide is a very orange red
(see fig. A.12). When it is mixed subtractively with phthalo blue, a near-
neutral, rather muddy color results. However, the additive mixture is a
beautiful purple.
Additive mixtures are often demonstrated by means of a
disk that can be rotated rapidly. It consists of color segments that can
be changed in size to alter the relative amounts of colors to be mixed as
the disk is rotated. When it is rotated rapidly so that the individual
90
80
70

60
g
<IJ
v
c
<d
tJ
50
<IJ
;:;:::


c
<IJ
v
(l;
40
"-
30
20
10
400
y 0
2% red oxide, 98% Ti02
0.5% phthalo blue, 99.5% Ti0 2
equal mixtures
2
, .... --- .... ,
, ....
, ..
I '
'
'
'
' :!>
,.
,.
--- subtractive } 1 % red oxide
0.25% phthalo blue
- - - - additive 98.75% Ti02
450 500 550
Wavelength (nm)
,.
,.
,.
,.
600
-----R----
------ ,.
,.
650 700
18 Chapter 2
colors cannot be discerned, the observer sees the additive mixture. Such
an instrument with a disk that has segmented color sections is called a
Maxwell disk, after James Clerk Maxwell, the scientist who demon-
strated the usefulness of the tool in studying color matching (Maxwell
1860; Judd and Wyszecki 1975). If the spectrophotometric curves of the
individual colors on the disk are known, and if the relative areas of each
are also known, then the spectral curve of the additive mixture can be
calculated, and the color notation also calculated.
Primary colors
We can define primary colors as sets of three specified colors from which
large numbers of colors and a neutral can be made by suitable mixture.
Primary colors for additive color mixture are not the same as the primary
colors of the dyes or pigments used in subtractive colorant mixture. For
additive color mixture, the three most useful primary colors are an
orange-red, a green, and a violet-blue, which are generally described
simply as red, green, and blue. We can observe how these colors mix by
illuminating a white wall or a projection screen with light from three
identical slide projectors, each equipped with a different colored filter
transmitting the color of one of the additive primaries. When the proper
amount of filtered light from each of the projectors is focused on the
same area of the white wall or screen, the resultant color is white light.
If the green light and the red light are focused on the same area, yellow
light results; green and violet mixed additively result in a greenish blue
(cyan); violet-blue and red mixed additively make a bluish red (magenta).
These colors (yellow, cyan, and magenta), which are the common sec-
ondary additive colors, are found to be the most useful primary colors in
subtractive colorant mixture. (In popular terminology, magenta, yellow,
and cyan are also referred to as red, yellow, and blue.) Thus in color
printing, the three primary colors of ink required for use in subtractive
colorant mixture are magenta, yellow, and cyan. We find the most famil-
iar use of the additive primaries in the orange-red, green, and violet-blue
phosphors in our television tubes.
These relationships are described in figure 2.3, which illus-
trates a Hue Circle of colors with primaries indicated on the circumfer-
ence of the circle. The additive primaries are marked with a plus sign (+),
the subtractive primaries with a minus sign (- ).The center, N, marks
neutral. A similar diagram is presented in Billmeyer and Saltzman (1981).
From this diagram one can see that there is no fundamental reason to
select a specific set of primary colors, either for additive color mixture or
for subtractive colorant mixture. For example, for additive mixture, the
three primary lights used for color matching may be obtained by filtering
white light through three appropriate filters as described earlier for addi-
tive mixture, or they may even be three monochromatic lights, that is,
three lights each of a single wavelength. Similarly, the subtractive pri-
maries may be any of a number of pigments or dyes in the hue ranges
of magenta, yellow, and cyan.
Two colors that result in a neutral color when mixed in
proper amounts are said to be complementary. In the next section, which
Figure 2.3
A color circle illustrating the relationships
of the additive color primaries, marked
with a plus sign (+), and the subtractive
primaries, marked with a minus sign (-).
Note that the additive and subtractive pri-
maries are complementary-i.e., 180
opposite on the color circle.
Colorimetry
red
+
N
blue-green
(cyan)
19
deals with the CIE system for describing color, the method for determin-
ing the exact hue of complementary colors in additive color mixture is
described. There is, however, no comparable, precise method for describ-
ing colorants that are complementary when mixed subtractively. In the
practical world of colorant mixture, the absorption and scattering of
light by colorants and by the color of the substrate affect the resultant
color of the admixture. That some colorants are essentially complemen-
tary is well known to the artist: mixing red paint with blue-green (cyan)
paint will result in a muddy gray paint; red and cyan are complementary.
These concepts of additive and subtractive color mixture and
of complementary colors in both types of mixture have importance
beyond the fact that each is the basis for a different type of colorimetry
system, that is, for an essentially different language for describing color.
When attempting to describe or identify the colors of objects, an under-
standing of these concepts is vital. If all objects were of uniform color,
the problem of measurement would be relatively straightforward. But
when the object is multicolored or patterned, the problem is not so easily
solved, and the analyst must be aware of the possibilities inherent in the
principles just described and of the pitfalls that may be encountered.
The little squares of yellow and blue described previously
are akin to the technique of the pointillists, the late-nineteenth-century
painters who discovered scientific additive mixture theory and, using
this principle, painted dots of relatively pure color to define their scene.
Since these paintings depend on the eye for color mixture, the viewing
20 Chapter 2
distance is all important. Viewed closely, the dots are visible in the
painting; at a distance, they are not, and we see the additive mixture. At
just the right distance, called the fusion point, they seem to "shimmer"
as the eye alternately integrates and separates them. (Museums should
display these paintings appropriately with regard for this fusion dis-
tance. ) The pointillist technique is an obvious attempt to use additive
color mixture as an artistic tool.
In most classical representative painting, this tool is also used
for artistic effect but without the obvious extensive use of dots based on
the strict application of the scientific theory. Most classical paintings, as
well as modern paintings done with certain traditional painterly style tech-
niques, are not uniform in color in every area. The artist may have dipped
the brush into several colorants before applying his paint to the canvas,
resulting in streaks of color. Or the paint may have been scumbled (result-
ing in a granular color surface achieved by rapid overpainting with a thin
layer of opaque or semiopaque color using a dry-brush technique), or
small highlights may have been applied. As the eye integrates (addi ti vely
combines) these mixtures, so will a color-measuring instrument, depending
on the size of the area viewed by the instrument. We must not forget this
possibility in the act of measurement; it is dealt with later at greater
length. For our purposes, the solution to minimizing the effect of additive
mixing of areas of different pigmentation on a painting is to make multiple
measurements, and this is practical with modern high-speed instruments.
For the skilled interpreter, spectral reflectance curves from a number of
areas will quickly indicate similarities that will provide a key to the basic
palette. This is part of the skill defined in this monograph.
The Cl E System
Many of the various descriptors for color, as it is perceived, are based
on what is known as the CIE system, from the initials of the Commis-
sion Internationale de l'Eclairage, or the International Commission on
Illumination, the organization that has administered the system since
its beginnings in 1931. (In older publications, ICI-the initials for the
English name of the commission-may be encountered. ) In the preceding
chapter on spectrophotometry, it was pointed out that, from an analysis
of spectral curve shapes, the color dimensions of hue, lightness, and satu-
ration could be deduced in a qualitative way. These three terms describe
the perceptual (that is, psychological) dimensions of color. In the intro-
ductory section to this chapter on colorimetry, it was emphasized that
there are only three basic methods for creating an orderly arrangement of
colors. The perceptual system, which is based on visual estimates of hue,
lightness, and saturation, is one of them. The CIE system is of the second
type, that is, a psychophysical system based on addi ti ve color mixture
and on the visual response to various wavelengths of light reflected or
transmitted to a normal human observer.
The colors that we see in a scene depend on the spectral dis-
tribution of the light illuminating it and on our spectral response to the
Colorimetry 21
light reaching our eyes after it is reflected or transmitted by the elements
in the scene. If these two aspects-the nature of the light source and the
response characteristics of the observer-can be defined, then the colors
reflected or transmitted in the scene can be described mathematically.
In 1931 the original method for the calculation of three-
dimensional color descriptors was introduced by the CIE, and it has been
accepted widely as an internationally recognized color language. Included
in the 1931 method were three standard light sources, indicated by the
letters A, B, and C, and the spectral color-matching functions for a stan-
dard observer. Later, the spectral power distributions of the light sources
were measured, and these data became the official definitions of illumi-
nants A, B, and C. In 1965 the CIE system was augmented by adding
the daylight, D, illuminant; the year before, a supplementary standard
observer had been added to the system. The 19 31 standard observer rep-
resents distance viewing based on a 2 subtended angle of viewing for
determining the color-matching functions; the 1964 standard observer
represents closer viewing, described as the 10 subtended angle response
or, roughly, as viewing at arm's length. For detailed information on the
CIE methods, the CIE and ASTM publications may be consulted (CIE
1986; ASTM Standard E 308), as well as any of the standard color sci-
ence texts listed in appendix E. The corresponding terminology is given
in the International Lighting Vocabulary (CIE 1987) and in ASTM
Standard E 284.
The 19 31 standard light sources were A, representing the
color temperature of an incandescent lamp at 2856 K; B, representing
mean noon sunlight obtained by filtering a tungsten lamp (4874 K); and
C, representing average daylight (6774 K). The last may be obtained by
using special blue filters over the source A incandescent lamp. Because
incandescent lamps do not include ultraviolet wavelengths to the extent
that sunlight does, all of these originally defined illuminants (sources)
were inadequate for measuring color as observed on fluorescent materi-
als, which depend at least partially on absorption of ultraviolet energy.
Recognizing this, the CIE in 1964 defined a series of daylight illuminants
that closely match the relative spectral power distribution of natural
daylight. Of these, the average daylight with 6500 K color temperature
(abbreviated D65) is the most important and most widely used. Figure
2.4 illustrates the relative spectral power distribution of the three origi-
nal CIE illuminants (A, B, C) and of D65. The large difference between
D65 and the others in the ultraviolet region 300- 400 nm is readily
apparent. Unfortunately, there is no real light source that simulates illu-
minant D65 closely, especially in the ultraviolet. Illuminant B has been
declared obsolete by the CIE. Illuminant C is still used, but D65 is
becoming increasingly more common and is the recommended illuminant
today. For nonfluorescent samples, illuminant C is adequate and, unless
otherwise specified, is assumed to be used in this monograph; it was used
in many colorimeters and in most of the older literature dealing with
nonfluorescent colorants. The other widely used illuminant is A, incan-
descent light. Used less frequently are fluorescent illuminants described
by the CIE (1986).
22
Figure 2.4
Relative spectral radiant power distribu-
tions of CIE illuminants A, B, C, and 065.
Note that illuminant B has been discon-
tinued as one of the standard illuminants
(from Wyszecki and Stiles 1982. Copyright
1982 John Wiley & Sons, Inc. Reprinted
by permission).
Figure 2.5
Diagram illustrating the method of addi-
tive tristimulus matching. The portion of
the screen showing the samples is
matched by varying the amounts of light
from the three-color projector illuminating
the other (match) portion of the screen.
Chapter 2
250
Q;
~ 150 1 - - ~ ~ ~ ~ - + - ~ ~ ~ ~ - + - ~ ~ ~ ~ - + - ~ - ~ ~ - - + ~ ~ ~ ~ ~ + - - ~ - I
0..
....
c
,,;
'6
~
<l.>
>
~ 100
Qi
0::
50
300 400 500 600 700 800
Wavelength (nm)
In addition to the nature of the light source, the second fac-
tor affecting the perception of the color of an object by an observer is
the spectral sensitivity of the eye. This visual response can be measured
in terms of the amounts of three primary lights required to be additively
mixed to match the light at each wavelength in the visible region of the
spectrum. A schematic diagram of the experimental method is shown in
figure 2.5. All humans with normal color vision do not have identical
responses. However, the differences are small in comparison to the differ-
Figure 2.6
Comparison of Cl E 2 and 10 standard
observer color-matching functions (from
Judd and Wyszecki 1963. Copyright
1963 John Wiley & Sons, Inc. Reprinted by
permission).
Colorimetry 23
ences that occur in persons with abnormal color vision (some degree of
color blindness). In 1931 the CIE established a standard observer in
terms of the average measured response characteristics for a group of
normal observers. Three response curves were defined, one for each
primary light used in matching many wavelengths across the spectrum.
The experiments were carried out using red, green, and blue primary
lights. For reasons of mathematical simplicity, the experimental data
were transformed to another set of primaries that bear no relation to any
real lights. These transformed response values, used ever since, are called
the color-matching functions, x, y, z, and are illustrated as spectral
curves in figure 2.6. When the mathematical transformations were made,
the Y tristimulus value was made to correspond to the lightness response.
The light incident upon the eye is the summation of the
product of the radiant power from the illuminant, S(A.), and the frac-
tion of this light that is reflected (or transmitted) by the object at each
wavelength, R(A.). This product summation is generally denoted as
ES(A,)R(A.)llA, where llA describes the wavelength interval used. The
wavelength interval, llA, is constant across the spectrum and must be
sufficiently small that the characteristics of the spectral curves are ade-
quately represented. Often llA is 10 nm, for example.
2.0
---+-- 2observer (CIE)
-0- 10observer (CIE)
1.5

::J
!ii
>
V>
::J
::J
E
1.0


400 500 600 700
Wavelength (nm)
24 Chapter 2
Knowing the spectral characteristics of the light reaching the
eye, a description of the color as seen by the standard observer can be
calculated. The product S(A.)R(A,) for each wavelength is multiplied by
each of the three color-matching functions, x, y, z, at the same wave-
length. The curves of the spectral color-matching functions for the 2
(1931) and for the 10 (1964) standard observers are illustrated in figure
2.6. The products are then summed (the symbol L: ) for each primary to
give a three-dimensional description of the color. The results are called
the tristimul us values: X, Y (lightness), and Z:
X = HS(A.) R(A.) x (A.) 6A.
Y = kLS(A.)R (A.) y (A.)6A.
Z = HS(A.)R(A.) z (A.)6A.
(2.1 )
where k is a normalizing factor to make Y equal to 100 for the white ref-
erence standard (the so-called perfect reflecting diffuser), thus describing
the lightness. The tristimulus values may vary depending on the number
of equally spaced wavelengths used in the summations. If t-,.A, is 1 nm or
less, the summations become true integrations (the symbol f ):
X = kfS(A.)R(A.) x (A.) dA.
Y = kfS(A.)R(A.)y (A.)dA.
Z = kfS(A.)R(A.) z (A.)dA.
(2.2)
This integrative process may be more clear if shown graphi-
cally, as in figure 2.7. The spectral curves for illuminant C and for the
reflectance of an object are shown at the top. From our knowledge of
spectrophotometry, it can be seen that the reflectance curve represents
that of a blue object of moderate lightness and purity (saturation). The
product of these curves, S(A. )R(A. ), which is shown immediately below the
illuminant and object curves in figure 2. 7, describes the spectral light that
would be incident on the eye. Below this curve are the spectral response
characteristics, x, y, z (color-matching functi ons ), for the 1931 stan-
dard observer. The product of multiplying each of the color-matching
functions by S(A.)R(A.) is shown by the curves at the bottom of figure 2.7.
When the relative areas under curves S(A.)R(A,) x (A.) , S(A. )R(A.) y (A.) , and
S(A.)R(A.) z (/..,) are determined (i.e., the curves are integrated), we find that
the tristimulus values for this blue object under illuminant C as seen by
the 1931 2 standard observer are X = 55.55, Y = 57.33, and Z = 90.25.
We can now state the results obtained with the CIE system.
This system provides methods for computing sets of tristimulus values
such that if two specimens have the same tristimulus values, for a stated
illuminant and observer, they match in color; if they do not have the same
tristimulus values, they do not match. Although, strictly speaking, the
19 31 CIE system was not designed to describe the colors or the differences
between them if they do not match, these applications and many more
described later have been added to the original system since its beginnings.
In the case of the tristimulus values described above, it can be
assumed that the color is blue because the Z tristimulus val ue is so much
larger than the X and Y values. However, if we had another blue sample
at similar lightness ( Y) but with a different reflectance curve, such that
X = 50.34, Y = 54.19, and Z = 85.69, it would be difficult to visualize
Colorimetry
100
90
80
70
60
s 50
40
30
20
10
Illuminant C
90
80
70
60
R 50
40
30
20
10
12.5% ultramarine
blue & 0.02 % black
mixture in rutile Ti02
400 450 500 550 600 650 700 400 450 500 550 600 650 700

100
90
80
70
60
SR 50
40
30
20
10
400 450 500 550 600 650 700
Wavelength (nm)
25

w w w
w w w
ro ro ro
w w w
x 50
40
30
20
10
y 50
40
30
20
10
z 50
40
30
20
10
400 450 500 550 600 650 700
Wavelength (nm)
400 450 500 550 600 650 700
Wavelength (nm)
400 450 500 550 600 650 700
Wavelength (nm)
I

90
80
70
60
90
80
70
60
I I

90
80
70
60
SRx 50
40
30
20
10
SRy 50
40
30
20
10
SRz 50
40
30
20
10
Figure 2.7
Graphical illustration of the calculation of
the CIE tristimulus values. The blue sample
is a mixture of 12.5% ultramarine blue,
0.02 % carbon black, and 87.48% rutile
Ti0
2
in a glossy paint medium.
400 450 500 550 600 650 700
Wavelength (nm)
400 450 500 550 600 650 700
Wavelength (nm)
400 450 500 550 600 650 700
Wavelength (nm)
the relationship between the two samples on the basis of the tristimulus
values, X, Y, and Z, alone. To aid in this visualization, the CIE has
defined chromaticity coordinates (symbols x, y, and z) derived from the
three tri stimulus values:
x
x= ----
X+Y+Z
y
y= ----
X+Y+Z
z= z =1-(x+y)
X+Y+Z
(2.3)
26 Chapter 2
To describe the color completely, the lightness (the Y tristimu-
lus value) must be included with two chromaticity coordinates, x and y,
to describe the three dimensions adequately. To avoid confusion, the tri-
stimulus values (X, Y, and Z ) are generally (except in very old literature )
multiplied by 100, such that Y for a perfect reflecting white diffuser
equals 100. The x and y chromaticity coordinates are designated as
decimal fractions.
We now have two ways to describe a color in the CIE system:
(1) in terms of its tristimulus values (X, Y, and Z), and (2) in terms of its
chromaticity coordinates, x and y, and Y.
Because the sum of all three chromaticity coordinates is unity,
two of them, generall y x and y, may be plotted on a two-dimensional
graph. When the chromaticity coordinates for each wavelength in the
spectrum are plotted, a horseshoe-shaped diagram is obtained, as illus-
trated by the outline in figure 2.8. This diagram is based on illuminant C
and the 1931 2 standard observer. The chromaticity diagram for illumi-
nant D65 is very similar.
The outside of the horseshoe diagram in figure 2.8, formed
by plotting the chromaticity coordinates, x and y, for the pure spectrum
colors, is called the spectrum locus and represents the purest and most
saturated colors that can occur. The short-wavelength violet is at bottom
left beginning at 380 nm; the long-wavelength red is at bottom right at
780 nm; and the green at 520 nm is at the top of the horseshoe. The
point C in the interior of the diagram is located at the chromaticity coor-
dinates for illuminant C and represents zero color. Approximate color
names associated with various parts of the diagram are indicated. These
are not sanctioned by the CIE, but they have been found to be helpful as
a guide in relating colors on the chromaticity diagram. The oval area
around the illuminant point, which has no color name, is the more neu-
tral area of the diagram. It includes hues decreasing in saturation as the
illuminant point (also known as the achromatic point) is approached.
This diagram represents the chromaticity plane. The third dimension,
which is the lightness described by the Y tristimulus value, is perpendicu-
lar to this plane.
Figure 2.9 shows another use of the chromaticity diagram.
One of the features of this diagram is an illustration of the additivity of
color mixtures. When any pure spectrum color is mixed with the illumi-
nant color, the resulting chromaticity coordinates fall on a straight line
connecting the point on the spectrum locus with the color of the illumi-
nant. The lines radiating from the illuminant point to the spectrum locus
are called lines of constant dominant wavelength, Ad, and describe the
hue in terms of the wavelengths in nanometers at the intersections of the
lines with the spectrum locus. The chromaticity of any two colored lights
mixed additively falls on a straight line connecting chromaticities of the
two lights. If the straight line passes through the illuminant point, the
two colors are said to be complementary. Thus an orange of 587 nm
dominant wavelength and a blue of 485 nm dominant wavelength are
complementary because their additive mixture line passes through the
illuminant point.
0.800
0.200
0.100
520 525
450
440
430
Colorimetry 27
0.000 ~ ~ ~ ~ ~ ~ ~
4 2

0
: : : : : ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~
0.000 0.100
400
0.200
Figure 2.8
Hue names associated with areas of the
1931 CIE chromaticity diagram (after
Judd 1950).
0.300 0.400 0.500 0.600 0.700
x
This property of additive mixture when plotted on a chro-
maticity diagram enables us to describe purples, which are not spectrum
colors, as being mixtures of red light (long wavelength) and violet light
(short wavelength). These nonspectral colors have no dominant wave-
length; they are described in terms of the colors to which they are com-
plementary. They are indicated on the diagram with a subscript c
following the dominant wavelength of the complementary green color.
The total distance from the illuminant point to the spectrum
locus represents 0-100% purity. Divided into percentage distances from
the illuminant, the series of concentric horseshoelike lines are a measure
of the excitation purity (or simply purity), Pe The series of smaller
28 Chapter 2
0.90 ~ ~
520
0.10
Figure 2.9
Chromaticity diagram for the 1931
standard observer. The section outlined
by the dashed square is enlarged in
figure 2.11.
0.20 0.30 0.40 0.50 0.60 0.70
x
Chromaticity diagram for illuminant C
horseshoes in figure 2.9 are constant purity loci. For any given sample
point on thi s diagram, the purity can be esti mated from precalculated
lines (Hardy 1936) or by direct measurement of the distance from t he
illuminant, divided by the total distance of the illuminant through the
sample point to the spectrum locus, and then multiplied by 100 to get
a percentage.
Thus, a third set of color descriptors in t he CIE system is the
dominant wavelength (/1.d), the excitation purity (Pe), and the lightness
(Y). Note that Y must always be included.
Table 2.1
Color descriptions in the CIE system: tri-
stimulus values (X, Y, and Z), chromaticity
coordinates (x and y), dominant wave-
length, and excitation purity.
Table 2.2
Chromaticity coordinates (x and y) and the
luminous reflectance, or lightness (Y), of
three blues.
Colorimetry 29
Dominant Excitation
x
y
z x y wavelength purity
31 .39 19.77 10.74 0.5071 0.3194 614.0 nm 56.6%
32.84 43.06 9.35 0.3852 0.5051 566.0 nm 69.2%
29.17 19.77 26.23 0.3880 0.2630 499.4c nm 36.0%
A few examples are given in table 2.1. The reader may
confirm the chromaticity coordinates and then use figure 2.9 to confirm
the dominant wavelength and purity.
The development of the CIE system has been only briefly cov-
ered; the reader is referred to the references in appendix E for more com-
plete explanations and for descriptions of the experimental procedures
and mathematical methods involved. It suffices to describe here the use-
fulness of plotting colors on a chromaticity diagram to obtain informa-
tion on the relationships of measured colors.
Examples
Figure 2.10 illustrates the spectrophotometric curves for three blue
paints. Blue 1 is made with ultramarine blue (12.5%), carbon black
(0.02%), and rutile Ti02 (87.48%); the curve shape at 600-700 nm is
characteristic of ultramarine blue (see fig. A.8 ). Blue 2 is made with
phthalocyanine blue (0.25%), carbon black (0.10%), and rutile Ti02
(99.65%); the curve shape at 600-700 nm is characteristic of phthalo
blue (see fig. A.10). Blue 3 is a 50:50 mixture of blue 1 and blue 2.
Note that the pigment concentrations are expressed as relative dry
pigment concentrations based on the fact that the pigment total is
100.00% regardless of the exact weights or volumes used. The chro-
maticity coordinates (x, y) and the luminous reflectance, or lightness
(Y), of the three blues are given in table 2.2. These values are for illu-
minant C and the 1931 standard observer. The Y tristimulus value
indicates that blue 1 is the lightest. But in what way do the three blue
paints differ in chromaticity?
By plotting them on an expanded portion of the blue section
of the chromaticity diagram, their relationship becomes clear. This is illus-
trated in figure 2.11, which is the area enclosed by the dashed square
shown in figure 2.9. Blue 1, made with ultramarine, is seen to be redder
(more violet) and lower in purity than blue 2, which is made with phthalo
blue. Blue 3 is midway between the other two, as it should be; it is a
50:50 mixture of the other two blues.
On the expanded portion of the chromaticity diagram
(fig. 2.11 ), the illuminant point as well as reference purity and dominant
wavelength lines are shown. Although helpful, they are not necessary for
Blue x y
y
1 0.2711 0.2800 53.33
2 0.2614 0.2827 50.19
3 0.2655 0.2813 51.48
30
Figure 2.10
Spectrophotometric curves for paints made
up of mixtures of ultramarine blue and
black (curve 1) and for phthalo blue and
black (curve 2), both with rutile Ti0
2
;
curve 3 is their mixture in equal parts by
weight.
Chapter 2
90
80
70
]
60
g
"'
u
c
cu
ti
50
"'
""

c
"'
"'
40 C1.
30
20
10
400
y 0 -----R----
blue mixtures in rutile Ti0
2
,,,,..-.........
. '."-
t ---- .
/ "'
h
....... __ ,
1 --12.5% ultramarine blue+ 0.02 % black
2 - - - 0.25% phthalo blue+ 0.10% black
3 --- 6.25% ultramarine blue+ 0.125% phthalo
blue+ 0.06% black
450 500 550
Wavelength (nm)
600
-
650 700
quick analytical work. With a little practice one learns to relate adjacent
chromaticity points qualitatively without the reference values, from the
visualization of the relationships of colors in the various areas of the
diagram. An atlas of expanded diagrams may be found in the Handbook
of Colorimetry (Hardy 1936). These diagrams can be used to determine
the dominant wavelength and the excitation purity for any set of chro-
maticity coordinates. Today the dominant wavelength and purity can be
calculated with a computer (McCarley, Green, and Horowitz 1965). The
computer programs supplied with many color-measuring instruments
calculate the chromaticity coordinates and also display them, and may
print them on a chromaticity diagram.
From graphing the chromaticity coordinates of the three
blues used in the example, we have determined that the first blue is red-
der and lower in purity than the others, and the values of Y indicate it is
lighter than the second blue. This is a qualitative description of what one
Figure 2.11
Enlarged section of the chromaticity dia-
gram (fig. 2.9) showing the locations of
the samples whose spectral curves are
illustrated in figure 2.10. Note the near-
straight-line relationship of the chromatici-
ties of the three colors.
Colorimetry 31
0.31
0.30
0.29
0.28
0.27
0.26
0.25 L - - - - - - - - - - - - - ~ - - J __ ..r:::.___J, ___ __J, ___ __J, ___ _J
0.25 0.26 0.27 0.28 0.29 0.30 0.31 0.32
x
would see. But the CIE chromaticity diagram is based on additive mix-
ture, and we are primarily interested in pigment mixtures, that is, sub-
tractive mixture. Of what use in subtractive mixture is this diagram,
which is based on additive mixture?
Subtractive mixtures do not fall on straight lines on the chro-
maticity diagram; some lines are relatively close to being straight, and
others do loops and turnarounds when the individual pigments are mixed
with white pigment. Figure 2.12 illustrates some typical pigment mixtures
at various ratios of rutile Ti02 The reasons why some lines loop around
and others approach straight lines are presented in chapters 3 and 4 in the
discussions of subtractive mixture and its dependence on the absorption
and scattering characteristics of the pigments used in the mixture.
If the area outlined by the dashed square on the chromaticity
diagram in figure 2.9 is expanded 100 times, as was done to create figure
2.11, the assumption of a straight-line relationship in subtractive mix-
tures that are very similar in color is not unrealistic. For example, if we
make a 50:50 mixture of blue 1 and blue 2-which is represented by
blue 3-and plot the chromaticity coordinates on the expanded diagram,
it is seen that the chromaticity of the blue 3 mixture in figure 2.11 falls
very close to the straight line connecting the chromaticities of blue 1 and
blue 2. Thus the assumption of a straight-line relationship between the
chromaticity points for similar colors is a practical and meaningful way
to interpret relationships of closely related colors made by subtractive
mixture. For example, the chromaticity coordinates and luminous
32
520
0.8
0.7
0.6
0.5
>.
0.4
0.3
0.2
0.1
0.1 0.2
Figure2.12
Loci of the chromaticity coordinates of
common pigments mixed with varying
amounts of rutile Ti0
2
white in a glossy
paint medium. Discontinuities are due to
changes in pigment volume concentration
(PVC). Dotted sections are interpolated
(Johnston 1973).
Chapter 2
1. phthalocyanine blue
2. indanthrone blue
3. carbazole dioxazine violet
4. quinacridone magenta
5. lithol rubine (transparent)
6. BON red
7. molybdate orange
8. vat orange
9. Flavanthrone Yel low
10. light chrome yel low
11. Yellow FGL (monoazo)
12. chromium oxide green
13. phthalocyanine green
700
0.3 0.4 0.5 0.6 0.7 0.8
x
reflectance, or lightness, Y, for blue 3 are very close to the average for
blue 1 and blue 2, as shown in table 2.3.
Although the pigmentations of blue 1 and blue 2 are differ-
ent, the calculated average chromaticit y is still close to that actually mea-
sured. If the target or aim color is blue 3, and blues on different sides of
the aim point on the chromaticity diagram had been prepared, a very
good estimate for a pigment formula that is a close match can be made.
No computer color-matching program would be needed to adjust the
composition to make a desired color. Even with computer color-matching
Table 2.3
Comparison of chromaticity coordinates
(x and y) and the luminous reflectance, or
lightness (Y), for blue 3 and the average
values for blues 1 and 2.
Colorimetry
Blue 3
Average of blues 1 and 2
x
0.2655
0.2662
y
0.2813
0 .2814
y
5148
51.76
33
facilities available, the technique described serves as an excellent evalua-
tion and guide.
The major topic of this monograph is the use of spectrophoto-
metric curves as a guide to pigment identification. However, as was
shown, the application of colorimetry is also indispensable in utilizing
visible spectral curve measurements to obtain maximum useful informa-
tion concerning the color composition of objects and materials of interest.
Color-Difference Equations
In the previous section, the CIE system was described as based on
additive color mixture. The system was not intended to be visually
uniform, such that a fixed distance in color space would represent a
constant degree of perceived difference in all regions of color space.
Consequently, since the beginning of the CIE in 1931, many attempts
have been made to transform mathematically the CIE notation into a
system that is visually more uniform. Many equations have been pro-
posed. For a history of this evolution, consult one of the basic reference
books on color in appendix E, or see the concise historical summary
given by Johnston (1973).
Suffice it to say that any unit of color difference can be help-
ful in describing the relationship between similar colors, but that the
ideal transformation into a truly uniform color space is still elusive and
probably does not exist. Nevertheless, whatever transformation equa-
tions are used, the resulting information can be helpful for compar-
ing similar colors. However, the magnitudes of the visual differences
between colors in different regions of color space are not generally com-
parable: one unit of color difference in the blue region does not repre-
sent the same perceived difference as one unit in the red region. To add
to the uncertainty in the reliability and usefulness of color-difference
measurements in the practical world, surrounding colors in a scene or
setting can have a profound effect on a perceived color and hence on
perceived color differences.
The reasons for this stem from the complexity of the human
perception process, involving the eye and the brain. Psychological effects of
contrast, adaptation, scene complexity, memory, association, and so forth,
all contribute to our perception. Physical differences in materials-gloss
and texture, for example-also introduce confusion. It would be all but
impossible to include factors for all of these visual variables in a single
color-difference equation, even if they were much better understood.
For some wonderful visual illustrations of these psychological
influences on color relationships, the reader is referred to any of the
34 Chapter 2
books by Ralph Evans (1948, 1959, 1974). Burnham, Hanes, and
Bartleson (1963) present some wonderful illustrations of contrast effects.
To those concerned with complex colored objects-paintings,
textiles, etc.-the psychological effects are extremely important. For
example, the author had the privilege of measuring the color used in
some very old Russian icons that appeared to be brilliantly colored.
Measurement revealed, however, that the colors were not brilliant in
terms of luminosity or purity; the color contrasts and juxtapositions
made them appear so. Likewise, measuring the indi vidual colors used in
some oriental rugs also revealed that the colors were not brilliant. Skilled
artists and artisans have always learned to utilize the psychological
effects inherent in color relationships to create their works of beauty. To
the person measuring the colors in these complex scenes, their true physi-
cal description can be a bit of a shock. For that very reason, objective
measurements are of great help.
In color-difference equations, the total color difference is
generally expressed as t.. E, the difference due only to chromaticity dif-
ferences as t..C, and the difference in lightness as t..L. The magnitude
of a unit of t..E calculated for the same pair of colors using different
equations is not the same and can vary widely. There is a common mis-
conception that a t..E of 1.0 represents a just perceptible difference.
Instead, for example, the t..E calculated from equation 2.4, which fol -
lows, represents, on the average, about three times a just perceptible
difference. This equation, which is one of two recommended by the
CIE, is formall y referred to as the 1976 CIE L''a "b''. equation, but it
is also known by its official CIE abbreviation: CIELAB (pronounced
"sea-lab" ). The CIE did not recommend it because it is the best color-
difference equation, but because it could be computed easily with a
desk calculator and because a single standard of usage would help to
decrease the confusion and misunderstandings that have existed from
the use of many different formulas. Its use is recommended for object
colors, particularly because it is scaled directl y to the ANLAB 40 equa-
tion (Adams-Nickerson equation with a scaling constant of 40) widely
used by industries dealing primarily with reflecting materials. Color
differences calculated with the ANLAB 40 equation may be converted
to the CIELAB basis by multiplying by 1.1 (CIE 1986). This is the
only instance in which a constant factor can be used to convert color
differences calculated using one formula to those obtained using some
other equation.
Industries for which additive color mixtures of lights is the
major concern prefer to use the 1976 CIE L''.U''.V''. (CIELUV, pro-
nounced "sea-love") equation because of the additivity of its accom-
panying u, v uniform chromaticity diagram. Note that CIELAB is not
similarly additive and does not have chromaticity coordinates or a chro-
maticity diagram. Texts listed in appendix E and written after about
1978 may be consulted for details about these equations.
In the following CIE L ,,. a,,. b ,,. formula, t..E ,,. represents the
total color difference; t..L ,,. represents the lightness-darkness difference;
t..a'' describes the redness-greenness (+red-green); and t..b" describes the
Table 2.4
Calculations showing color differences
between blue 1 and blue 2.
Blue x
y
51.63 53.33
2 46.42 50.19
z
Colorimetry
yellowness-blueness (+yellow-blue) (CIE 1986; ASTM Standards D
2244, E 308):
35
(2.4)
where the ~ sign refers to the difference between the sample and the
standard in the three dimensions: L (lightness); a (redness-greenness);
and b (yellowness-blueness), and
L''. = l16(Y/Y
0
)
113
- 16
a''. = 500[(X/X
0
)
113
- ( Y/Y
0
)
113
]
b'' = 200[(Y/Y0 )
113
- (Z/20 )
113
]
(2.5)
(2.6)
(2.7)
The subscript n refers to the tristimulus values for a per-
fect reflecting diffuser for the illuminant and observer used. For
illuminant C and the 2 standard observer, Xn = 98.074, Yn = 100.000,
and Zn= 118.232.
The above equations are correct when XIX' Y/Y' and Z/Zn
are all greater than about 0.01, as is usually the case with object colors.
The CIE (1986) and most reference texts written after 1978 provide
alternative equations for smaller values.
For judging the direction of color differences between two col-
ors, the CIE 1976 hue angle, hab, and chroma, C'' ab, may be calculated by
(2.8)
(2.9)
When describing the difference between two colors using the equation
for the total color difference, ~ E ,,. ab, the lightness difference and the
chroma difference are calculated by
tiL" = v (sample) - L ,,. (standard) (2.10)
and the chroma difference is described as
tiC" = c: (sample) - c: (standard) (2.11)
Note that the hue angle, hab, is expressed in degrees (eq. 2.8)
and cannot be used in any equation to describe the total color difference,
~ E It is useful, however, for visualizing the relationship of colors using
graphical techniques, particularly for purposes of hue specifications.
The contribution of the hue difference, as distinct from the
chroma and lightness differences, can be computed from the total color
difference (eq. 2.4) by
(2.12)
Table 2.4 is an example of the calculation for blue 1 and blue
2, whose spectral curves are illustrated in figure 2.10.
L* a* b*
c ab
85.52 78.07 -1.64 -17.46 17.54 84.63 + 180 = 264.63
80.96 76.19 -7.60 -17.45 19.03 66.46 + 180 = 246.46
36
Figure 2.13
Illustration of a method of plotting CIE
a*b* coordinates. The small circles labeled
1 and 2 are calculated coordinates for the
curves of the two blues illustrated in figure
2 .10. It can be seen that they fall in the
quadrant bounded by hue angles of 180
and 270.
Chapter 2
CIE a* b* diagram
T T
T
T T
h = 90
+2 I- -i +2
+1 I-
+b
---1 +1
+a h = 0(360)
0 l--------------
1
-++<j1---------------I 0
h = 1so -a
-1 I-
-b
---1 -1
0 0
-2 I- 2 ---1 -2
h = 270
l l l l l
-2 -1 0 +1 +2
A graph of a''. versus b''. is illustrated in figure 2.13. (Because
the color is in the third quadrant, 180 have to be added to the calculated
hue angle as indicated.) It is apparent that blue 1 is redder than blue 2.
The color differences calculated for blue 2 versus blue 1 as
standard are
!::.L '
-1. 88
!::.a''"
-5.96
!::.b''"
-0.01
!'iH''" ab
5.08
From these values it can be seen that the major difference between these
two blues is in hue (t... H): blue 2 is much less red (or is greener) than blue
1 (!:la''. is very negative). Blue 2 is also darker (!:'>.L''. = -1.88 ).
Most modern computerized color-measuring instruments use
the CIELAB equation, not because it is the best-modifications to it have
been made that supposedly better fit visual data-but because, in support
of the CIE recommendations for standard practice, the equation is widely
recognized and understood. Suffice it to say: whatever unit of color dif-
ference is used in a research report, the equation used must be specified
for the benefit of the reader. And for quality control applications, the
specific unit to be used must be agreed upon by the correspondents con-
cerned. The CIELAB color-difference equation (eq. 2.4) will be used
throughout this monograph.
It has been common, in specifying total color difference, to
refer to the so-called National Bureau of Standards (NBS) color-
Colorimetry 37
difference unit. This NBS unit, l'lE, is calculated according to the follow-
ing equation:
where f g takes account of the masking influence of a glossy surface on
the detection of color differences, k relates the lightness and chromaticity
scales, and a and ~ a r e defined below. The gloss factor is f g = YIY + K),
where K is usually taken as 2.5, and k is usually taken as 10.
2.4266x - l.363 ly-0.3214
a = ~ ~ ~ ~ ~ ~
1.0000x + 2.2633y + 1.1054
~ = 0.5710x+l.2447y-0.5708
l.OOOOx + 2.2633y + 1.1054
Regardless of how it is obtained, the only true NBS unit is
that calculated according to the above equation, which is Hunter's
modification of Judd's original equation (Judd 1939; Hunter 1942).
When this equation was scaled, the assumption was made that a color
difference of 1.0 l'lE was a good commercial color match, even though
it was known to be two to four times the just perceptible difference.
The general feeling was that this was a reasonable limit to which com-
mercial color matching could be controlled. Unfortunately, there has long
been a misconception that 1 NBS unit or that 1 l'lE''-.b-or another equa-
tion scaled to the NBS unit-is a just perceptible difference. This is not
correct. This "true" NBS equation was never used widely, probably
because it was too complicated for the equipment that was readily
available at that time.
Many attempts have been made to transform the CIE color
space into more uniform color spaces and better color-difference equa-
tions, with better meaning in terms of the ease of calculation and mea-
surement or in terms of better agreement with visual evaluation. Some of
these equations were scaled, on average, to agree in magnitude with the
NBS unit and were then, unfortunately, called "NBS units." For example,
the constant 40 in the ANLAB 40 color-difference equation mentioned
above (in the discussion of the CIELAB equation) was added as a scalar
constant to adjust the magnitude to be close to the NBS unit, l'lE. As a
result, one must always ask which equation was used when one encoun-
ters references to an "NBS unit of color difference." In the CIELAB
equation presented earlier, l'lE is also scaled to agree generally in magni-
tude with the NBS unit, but it is not an NBS unit either.
One of the notably different color-difference equations used
in the more recent past was based on the research of MacAdam (1942,
1943; Brown and MacAdam 1957). His work resulted in color-difference
equations (FMC equations) based on perceptibility ellipses obtained by
matching lights (Chickering 1967, 1971; see also ASTM Standard D
2244). These equations were scaled so that 1 unit of difference, !le, is
equivalent to a just perceptible difference. (The unit scaled to the just per-
ceptible difference is designated !le here to distinguish it from l'lE, which
is scaled to the NBS unit.) A difference of 1.0 !'le based on MacAdam's
38 Chapter 2
perceptibility thresholds is on the average 0.2 to 0.5 in NBS-scaled units.
No fixed conversion factor exists, however, because of the very different
mathematical approaches used. The reader is directed to the standard
texts on color science in appendix E for further information on the
numerous color-difference equations and their supposed merits.
Since the adoption of the 1976 CIE recommendations, a num-
ber of new color-difference equations have been developed. Most promi-
nent among them is the CMC equation (McDonald 1980), which is
based on the L,C, H form of CIELAB and has significantly improved
visual uniformity.
Tristimulus filter colorimeters
Color-difference equations were originally developed for industrial quality
control. In the 1940s and the 1950s, photoelectric spectrophotometers
were either manually operated and hence very time consuming to use, or
automatically recording and consequently very expensive. This situation
led to the development of tristimulus filter colorimeters designed to mea-
sure how closel y a sample matched a standard. Hence they were often
called color-difference meters. Their design incorporated sets of three
filters that provided measurements approximating the three tristimulus
values X, Y, Z. The filters were selected to take into account the spectral
characteristics of the light source and of the detector response, in addition
to the x, y, z color-matching functions of the CIE 1931 standard
observer and illuminant C. Thus the instruments were an attempt to per-
form, in essence, an optical integration instead of a mathematical integra-
tion at individual wavelengths to obtain the tristimulus values X, Y, and
Z. The readings so obtained were called R, G, and B, and were in terms
of percentages. To obtain the tristimulus values for the 1931 standard
observer and illuminant C, the following equations were used:
X = (0.8 R + 0.2B)/0.98
Y=G
Z = 1.18B
(2.13)
Later instruments used a fourth filter to measure the blue loop in the X
tristimulus value.
A number of tristimulus filter colorimeters were designed.
Most of them utilized bidirectional geometry (45/0 or the reverse).
Later, instruments were developed that utilized an integrating sphere.
Widely used color-difference meters incorporated a simple
transformation of R, G, and B into a color-difference equation (Hunter
1942, 1958) as follows :
L = 10G
112
al = 17.S(R - G)/G
112
bl= 7.0(G - B)IG
112
2 2 112
tiE = [(L'iaL) + (tibL) + (tiL)2]
(2.14a )
(2. 14b)
(2.14c)
(2.1 5)
This simple equation represented an attempt to approximate the NBS
color-difference equation. Because the instrument was inexpensive and
simple to operate, it was widely used for quality control in paint-related
Colorimetry 39
industries and for other applications. The total-color-difference unit, t..E,
was widely called an NBS unit. However, the equation built into the
Hunter Color Difference Meter is not the NBS equation, and the t..E
obtained should not be called an NBS unit, although it is scaled to the
true NBS unit. Unfortunately, the t..E determined by use of the above
equation was often called an NBS unit of color difference, contributing
considerably to the confusion over the equation actually used.
Tristimulus filter colorimeters were not designed to measure
tristimulus values accurately. Indeed, the results from early instruments
varied significantly from values obtained by spectral integrations; they
varied also from instrument to instrument. Later instruments were
more carefully controlled as improvements in technology developed.
Tristimulus filter colorimeters play a useful role, but only if their limita-
tions as well as their advantages are properly understood. They were
designed to compare materials with similar composition, that is, similar
spectral curve shapes, and with similar appearances, such as surface
reflection and uniformity.
The use of an optical integrating technique (filters) to obtain
color differences has a very real advantage over the use of abridged spec-
trophotometry (measurements at a limited number of wavelengths, such
as 20 nm intervals) or abridged calculations of spectral data (such as cal-
culations of tristimulus values at 20 or even 10 nm). This advantage
appears when measuring high-chroma (very saturated) yellows, oranges,
and reds (also purples). The difference in the curves of two very similar
colors of this type is generally manifest in the spectral region where the
curves begin their steep rise. The difference may be over a few nano-
meters (5, for example) in width but would be missed by the use of a
limited number of spectral points. And a color match would be indi-
cated, whereas visual examination would indicate no match.
Additional comments
Distinctive color-difference specifications are often written that depend
on ( 1) the area of color space involved-the green, red, or blue area, etc.;
(2) the acceptability of variations in specific materials; and (3) prefer-
ences. If the color of butter is on the green side of standard, even by a
small amount, it is unacceptable. However, butter that is more golden
in color, even if different from the standard by the same color difference
or more, is perfectly acceptable. Thus, the type or direction of the color
difference from the standard is often more important in "acceptability"
than is the total color difference, t..E. Experience has shown that dif-
ferences in hue are most important; differences in saturation are
less important; and differences in lightness are least important. The
ratio of difference of tolerance has been estimated to be hue 3, satura-
tion 2, lightness 1. However, all such generalities can fail depending on
the color composition in a complex scene and on preferential directions
of color differences that may be considered tolerable or acceptable.
The use of color-difference equations to describe color change
resulting from exposure to deleterious conditions over a period of time
requires special interpretation; this includes, most importantly, analysis
40 Chapter 2
of spectral curve changes resulting from changes in colorant concentra-
tion (see chap. 4). Simply reporting changes in terms of t...E is almost
useless to the researcher, regardless of the equation used, because quan-
tifying exposure-related changes involves making measurements over
extended periods of time. The reliability of such measurements must be
ascertained: How carefully is the instrument calibrated and operated?
What is the reproducibility with which the instrumental measurements
can be made? Are the measured values greater than the uncertainties with
which the measurements can be reproduced over the period of time?
Consider also the nature of the change: How uniform is it? Has the gloss
changed? Has colorant been lost? Have all the colorants in a mixture
decreased in concentration in a similar way? Has a paint chalked rather
than faded? From a simple t...E color-difference figure, little of this infor-
mation can be obtained that will contribute to an understanding of the
changes that have taken place. The directional components of color dif-
ferences, obtained by plotting and comparing colors on a chromaticity
or CIELAB a''"b''. diagram, can be more informative. However, when
describing changes following exposure in analytical terms, colorant con-
centration calculations described in chapter 4 and study of the spectral
curve changes can be more meaningful in learning the source of, or rea-
sons for, the observed visual change.
Metamerism
Metamerism is a phenomenon that plagues all those who must match
colors. Two colors are said to be metameric if they appear to be the same
color (that is, they match) to an observer under one set of illumination
and viewing conditions, but appear to be different if any one of these
conditions is changed. The conditions include the spectral character of
the illuminant under which the pair of colors is observed; the spectral
reflectance or transmittance of the objects; the particular response of the
observers looking at the color pair; and the directional or geometric con-
ditions of the viewing.
Illuminant differences are illustrated by the spectral distri bu-
tions of daylight (bluish), incandescent light (orangish), and fluorescent
lamps (with various phosphors and strengths of the mercury lines).
Observer response differences exist among persons with normal color
vision. Most of us are not "standard observers" as represented by the
average figures used in the CIE calculations. For a detai led analysis of the
uncertainties, the papers by Nimeroff (1957, 1966) may be consulted.
Allen (1970) described a calculation for observer metamerism. Older per-
sons whose lenses have yellowed to some degree do not see metameric
pairs the same as younger people. Viewers with any degree of abnormal
color vision should not be involved in making color-match judgments. It
is advisable that all persons involved in making color-match decisions be
checked for color vision deficiencies.
When object colors that match-to one person in one light
source-no longer appear to be identical if either the illuminant is
Colorimetry
changed or a different observer looks at them under the same light, the
mismatch is the consequence of differences in the spectral distribution,
that is, the reflection or transmission curves, of the two object colors.
The two object colors are metameric as disclosed by the measurement
of their different spectrophotometric curves. One can modify the term
metamerism to describe the particular change made in viewing condi-
tions that led to the observation of the phenomenon. Thus illuminant
metamerism describes the color shift observed when the light source is
changed, and observer metamerism describes the change that occurs
when a different observer views the pair of colors. In either case, the
color changes result from differences in the spectral distribution of the
pair of samples being observed.
41
A pair of colors that seem the same when viewed under one
angular condition but no longer seem to match when the angle of illu-
mination or viewing is changed have been called geometric metamers,
and the phenomenon called geometric metamerism (Johnston 1967a;
FSCT 1995; Pierce and Marcus 1994). In this case, the change in color
relationship observed with a change in angle of illumination or view is
generally attributable to differences in the structure of the pair of
objects or materials being compared, that is, to differences in the surface
characteristics or in the internal particle composition or distribution.
Such pairs may also exhibit differences in their spectral distribution
curves due to variations in the types of chromatic pigments used so that
changes in the spectral character of the illumination and in the response
characteristics of the observer will also occur, resulting in two types of
metameric differences.
In the paragraphs that follow, illuminant metamerism and
observer metamerism will be discussed before geometric metamerism. Use
of the term geometric metamerism for color changes in color pairs that
match under one set of angular viewing conditions but do not match if
the angular conditions are changed has been questioned. This issue will
be discussed at the end of this section.
An excellent paper on metamerism due to pigment spectral-
curve differences, and its importance in the retouching of paintings, is
that of Staniforth (1985). One of the interesting facts she points out is
that metameric pairs, which seem to match under a particular light
source and to most observers, may look different if photographed in
color under the same light source. This occurs because the three color-
sensitive layers in color film used to produce the colored image are differ-
ent from the normal sensitivities of the eye, particularly the red-sensitive
film layer. The film reacts differently from the eye to differences in spec-
tral distribution. In general, the differences in sensitivity in various color
films are much larger than those among normal human observers. One
can consider this film sensitivity difference as a special case of observer
metamerism, the color film being the observer.
In her article, Staniforth shows the spectral reflectance curves
of metameric matches to traditional blue pigments, using a variety of
blue pigments available today. In each case only one blue pigment was
used in combination with other colored pigments as needed to make the
42
Table 2.5
Metameric matches to blue 1 and blue 2
in fig. 2.10 (CIE notation for illuminant C,
1931 standard observer).
Chapter 2
match. Closer, less metameric matches could probably have been made if
two or more blue pigments were used in combination. Such complex for-
mulations would be difficult to make by visual means alone, however.
The blues illustrated in figure 2.10 are not metamers because
they would not match under any normal light source or to any normal
observer. The definition of metamerism bears the stipulation that two col-
ors match under some condition of illumination and viewing. The same
two blue pigments used to make blues 1 and 2 illustrated in figure 2.10
could be used in a formulation that would match the other under one set
of standard conditions. For example, a match to blue 1, made with ultra-
marine blue, was calculated using blue 2, phthalo blue; a match to blue 2,
made with phthalo blue, was calculated using blue 1, ultramarine blue.
Table 2.5 shows the formulations obtained and the tristimulus values and
chromaticity coordinates for these mixtures. The tristimulus values are
based on 5 nm interval reflectances from 400 nm to 700 nm. The data are
calculated for illuminant C and the 1931 standard observer, and measure-
ments made with the specular reflection component were excluded.
The tristimulus values and chromaticity coordinates indicate
that the metameric matches are not perfect, but nonetheless they are very
close. The calculated color differences using the CIELAB equation (eq.
2.4) suggest that the differences are probably imperceptible: t.E''.ab = 0.29
for the first pair, and t.E''. ab = 0.24 for the second pair. Figures 2.14 and
2.15 illustrate the spectral reflectance curves for the pairs. Comparing
the areas between the curves would lead one to suspect that the blue pair
2-2a would be more metameric than pair 1-la because there is more
space between the curves.
The use of the term "more metameric" implies that there is a
degree of metamerism, that is, some pairs are more severely metameric
than others. In 1972 the CIE published " Special Metamerism Index:
Change of Illuminants" (CIE 1972, 1986). Assuming a pair of colors is a
perfect match under a particular illuminant, the color difference-t.E-
calculated for another, defined test illuminant constitutes the Metamerism
Index. An example is the calculation of the color difference in illuminant
Blue Pigmentation x
y
z x y
12.5% ultramarine blue 51.63 53.33 85.52 0.2711 0.2800
0.02 % black
87.48% Ti0
2
1a 0.172 % phthalo blue 51.52 53.31 85.22 0.271 1 0.2805
0.181 % Quinacridone Red
0.014% black
99.633 % Ti0
2
2 0.25% phthalo blue 46.42 50.1 9 80.96 0.2614 0. 2826
0.10% black
99.65% Ti0
2
2a 14.435% ultramarine blue 46.24 49.93 80.57 0.2616 0.2825
0.158 % phthalo greenb
85.407% Ti0
2
'See fi g. A.16 in appendix A.
bSee fi g. A. 22 in appendix A.
Figure 2.14
Metameric match (curve 1 a) to blue 1
made with the same blue pigment, phthalo
blue, used in blue 2 (fig. 2.10). Note that
Quinacridone Red had to be added to
achieve the necessary redness.
Colorimetry 43
A for a metameric pair that is a perfect match in illuminant C. The index
would be the t..E determined in illuminant A. Color differences under
different illuminants are routinely used in computer colorant formula-
tion calculations, but many workers have found this to be of limited
value in the real world when the samples are prepared and do not
match precisely.
A more general index of metamerism was described by
Nimeroff and Yurow (1965). The spectral differences between the curves
of a pair of metamers were weighted by the observer color-matching func-
tions and the spectral power distributions for the illuminant in which the
pair match. Their suggested index, however, has not been used extensively.
In the real world it is seldom the case that a pair of metamers
matches perfectly to the standard observer under the calculated illumi-
nant. However, if the match is very close, as is the case for the blues
described above, calculating the color difference for a different illumi-
nant can be helpful. Thus, the first pair of blues (blues 1 and la in
90
80
70
<il 60
g
"'
u
c
.s
50
;:;:::

.....
c
"' i:::'
g: 40
30
20
10
400
y 0
blue mi xtures in rutile Ti0
2
--- 12.5% ultramarine blue+ 0.02% black
- - - 0.172% phthalo blue+ 0.181 %
Quinacridone Red+ 0.014 % black
450 500 550
Wavelength (nm)
600
14---- R----
650 700
44
Figure 2.15
Metameric match (curve 2a) to blue 2
made with the same blue pigment, ultra-
marine blue, used in blue 1 (fig. 2.10).
Note that phthalo green had to be added
to achieve the necessary greenness.
Chapter 2
fig. 2.14), which had a total color difference of 0.29 in illuminant
C, has a in illuminant A of 0.61. The second pair (blues 2 and 2a
in fig. 2.15), which had a of 0.24 in illuminant C, has a of
0.41 in illuminant A. Thus, we see that the color differences for both are
larger in illuminant A. The differences would be readily perceptible in
illuminant A but would not be severe.
The CIE recommends that if an exact match cannot be made,
a formulation correction should be made. If that is not possible, "an
additive or multiplicative correction could be made" (Judd and Wyszecki
1975). While the CIE did not describe the method for doing this correc-
tion, Brockes (1969, 1970) has worked out the technique. He found, as
have other workers, that the multiplicative method gives the best agree-
ment with visual evaluations. Applying this method to our blue pairs
(table 2.5) serves to illustrate the procedure. Let the original samples,
1 and 2, be the standards, and la and 2a be the matches.
90
80
70
/
/
/
I
60
I
<a
I
g
I Q)
u
c
.:s
50
u
Q)
q::

.....,
c
Q)

Q)
40
CL
30
---
20
10
400 450
y 0
blue mixtures in rutile Ti0
2
0.25% phthalo blue + 0.1 % black
14.435% ultramarine blue+ 0.158%
phthalo green
500 550
Wavelength (nm)
600
-----R----
/
/
2a
650 700
Colorimetry
1. Determine the factors by which X, Y, and Z under the
reference illuminant (C in the case of the blues) must be
multiplied to give the exact values of X, Y, and Z for the
standard:
For pair 1: fx = 51.63/51.52 = 1.002135
fy = 53.33/53.31=1.0003752
fz = 85.52/85.22 = 1.0035203
45
Thus, if the tristimulus values of the metameric match (sample la in this
case) are multiplied by these factors, it will be a perfect match.
2. Multiply the tristimulus values of the metameric match in
the test illuminant (A in this instance) by these same fac-
tors to adjust them for the mismatch, assuming that the
match had been perfect under ill uminant C. The tristimu-
lus values under illuminant A for the standard in this case,
sample 1, are X = 53.27, Y = 50.90, Z = 25.63. The tri-
stimulus values for sample la in illuminant A are X =
53.05, y = 50.88, z = 25.40.
Xcorr = 1.00214 {53.05} = 53.16
Ycorr = 1.00038 {50.88} = 50.90
Zcorr = 1.00352 {25.40} = 25.49
Calculate the color difference of the adjusted tristimulus values between
sample l a and the standard (sample 1) in illuminant A. This is '1E''.ab =
0.425, which is smaller than the color difference (0.61) obtained without
making the correction. When the tristimulus values for the second pair are
similarly corrected, based on the degree of the mismatch in illuminant C,
the color difference in illuminant A is 0.788, which is much larger than
what the uncorrected data showed and is consistent with the larger areas
between the curves for the second pair of blues. In the experience of this
author, the index determined in this way is a better index of metamerism
for samples in the real world than an index without the correction.
Other factors make it difficult to establish a metamerism
index that would serve to describe the degree of metamerism in agree-
ment with visual observation. One, already mentioned, is the fact that
most of us are not standard observers. That is, our eyes' color-matching
functions (our visual responses) are not the same as the figures used for
the CIE standard observers. This fact always presents a problem when
one is evaluating the closeness of a metameric match that must satisfy
more than one observer. Two colors may appear to match when viewed
by the person who made the samples, but not by someone else, perhaps a
client, who evaluates the pair under the same illumination and viewing
conditions. This is the very practical meaning of observer metamerism.
An important factor that basically affects all CIE measure-
ments, and hence also all equations based on CIE values, is the method
of making the measurements, particularly the instrument band pass (reso-
lution) and the wavelength interval used in the measurement and in the
integration calculation. The color differences calculated by different
46 Chapter 2
laboratories using instruments of different design may be different. For
nonmetameric pairs, or those with very similar spectral curve shapes, the
use of large wavelength intervals, such as 20 nm (16 wavelengths ), wi ll
have a minimal effect on the color differences calculated, even though the
tristimulus values will not necessarily be the same as those based on, say,
5 nm intervals. However, when the samples are metameric and have
notably different curve shapes, the color differences calculated will be
very much dependent on the wavelength intervals used; the greater the
differences in area between spectral curve shapes of the individuals of the
metameric pair, the more serious the discrepancy may be in the calcu-
lated tristimulus values and, hence, in the calculated color differences.
A band pass and wavelength interval of 5 nm or less has been shown to
provide values that are not significantly different from those obtained by
continuous integration (CIE 1986; ASTM Standard E 308).
One may well ask: If a continuous integration is desirable,
why not use a tristimulus colorimeter to make the measurements, since
such instruments simulate a continuous integration by the use of filters?
The answer is that most of the tristimulus filter instruments available in
the past, and some that are still used today, were essentially color blind:
the combination of light source, filters, and detector response charac-
teristics did not match the CIE color-matching functions and illuminant
distribution very well. They were, and are, basically unsuitable for mea-
suring the color difference between colors that are metameric. Some tri-
stimulus colorimeters available today approach agreement with the CIE
functions for one illuminant and one observer, but they cannot provide
the measurements for another illuminant or observer necessary for deter-
mining the degree of metamerism.
An additional characteristic of the phenomenon of metamer-
ism, although perhaps obvious, is the fact that the degree of metamerism
can be most severe on low-chroma colors, such as near neutrals, which
can be made with many different colorants. Bright, high-chroma colors,
especially yellows, oranges, and reds, if they can be matched at all, wi ll
have to have almost identical spectral curves and will not be noticeably
metameric even if made with different colorants. On such colors, which
have very steep rises in reflectance, use of a band pass and measurement
interval of 5 nm or less is desirable for accurate measurements. Even
when using 10 nm intervals, one can miss the important wavelength
where the steep rise in the curve takes place. On such high-chroma
samples, which are essentially nonmetameric, as mentioned before, a tri-
stimulus filter colorimeter is more satisfactory for measuring small differ-
ences in color; it provides more sensitive and more accurate results than
those obtained with measurements at 20 nm or even 10 nm intervals. For
these high-chroma yellows, oranges, and reds, an index of metamerism is
generall y not necessary, but accurate measurements of color differences
are. The blue metamers described earlier were measured at 5 nm inter-
vals. If the degree of metamerism were determined by measurement at
20 nm intervals, the index would be different.
Formulation of nonmetameric matches for pastel shades and
near neutrals can be challenging without some information on the col-
Colorimetry 47
orants involved. This can be learned either from the spectral reflectance
curve shape or from some other analytical technique. Determining the
palette of colorants used by measuring the reflectance spectra in various
parts of a multicolored or variegated object can often be helpful. How-
ever, even though an identification has been achieved, decisions still often
must be made on whether or not to use the same colorants in formulat-
ing a match, taking into consideration such factors as their lightfastness,
availability, cost, and so forth.
A basic feature of the spectral curves of metameric pairs is
that they must intersect or cross one another at least three times (Stiles
and Wyszecki 1968). Our metameric blues intersected five times. A per-
fect curve match represents an infinite number of crossovers; a pair of
samples with curves that intersect only once or twice are not metamers.
Because metameric color matches cause so much confusion,
it is obviously desirable to avoid having to make them. Instead, it would
be better to use the same colorants in the match that are used in the
standard. This has traditionally been attempted by conservators when
inpainting losses in objects of fine art. However, this is not always
possible. So if imperfect matches must be settled for, it is best to make
them with as little metamerism as possible.
Geometric metamerism-a special case
Pairs of colors that appear to be the same at one angle of illumination
and a particular angle of viewing, but do not appear to be the same if
either the angle of illumination or the angle of viewing is changed, have
been called geometric metamers. The term geometric metamerism has
not been accepted officially by any standardizing organization (Johnston
1967a; FSCT 1995; Pierce and Marcus 1994). Nevertheless, it remains a
useful descriptor for this phenomenon. Such pairs of colors change from
a match to a mismatch as the illumination angle is changed while the
viewing angle is kept constant, and vice versa. The change in color,
lightness, and chromaticity can be observed by holding the sample pair
side by side in a viewing booth, observing the color from the front and
then rotating the sample pair with the result that the light from the
lamps in the inside top of the booth illuminates the samples at different
angles. If both colors appear to match at one particular angle of illumi-
nation, but then one color appears lighter if rotated in one direction and
darker if rotated in the other direction from the match point, or if their
hues or chromas appear to change, the pair is exhibiting geometr ic
metamerism.
The causes of the phenomenon are generally associated with
structural differences rather than spectral differences. The structural dif-
ferences may occur inside the materials, or they may occur on the sur-
face, particularly with materials of low gloss.
In order to measure the reflected color change with angle, a
goniospectrophotometer is necessary (Judd and Wyszecki 1975). Such an
instrument was described in chapter 1. All of the angular measurements
described in this monograph, including the ones in chapter 5 and in the
description of surface reflection in chapter 6, were made using the Trilac
48 Chapter 2
spectrophotometer operated in the gonio mode; the tristimulus val ues
were calculated at 10 nm wavelength intervals.
When colorimetric metamerism (metamerism due to differ-
ences in spectral curve shape) as well as geometric metamerism both
occur in a pair of samples, the problem can be very confusing and
difficult to explain without the aid of goniospectrophotometry.
Discussion
There are differences of opinion about the proper definitions for many
terms used to describe appearance phenomena such as metamerism. In
the author's opinion, the criteria for a good definition are that it be
(1) as descriptive of the particular appearance aspect as possible and
(2) that it be as simple and as easily understood as possible. The subject
of this section is metamerism, an appearance phenomenon, and its
definition given in the first paragraph of this section describes it without
reference to causes. Various causes of the phenomenon are described in
terms of changes in the characteristics of the illumination, observer, or
angular viewing conditions. Thus, we have used the terms illuminant
metamerism, observer metamerism, and geometric metamerism. To the
author, these terms are simple and understandable, and they will be used
throughout this book. If the specific type of metamerism is not indicated,
it is assumed that the phenomenon is caused by di fferences in the spec-
tral reflection or transmission curves of the materials being compared.
Other terms, which haven't gained general acceptance, have
been proposed to describe the color change in materials when the angu-
lar conditions of illumination or view are changed. For example, a single
color may appear different at different angles. Such a color has been
described as goniochromatic, and the phenomenon is called goniochro-
matism (Hemmendinger and Johnston 1970; FSCT 1995). It follows that
a pair of colors that appear to match under one angular arrangement of
light source and observer, but no longer match when the conditions are
changed, are called a goniometachromatic pair, and they are exhibiting
goniometachromatism. If a pair matches at all angular conditions, they
are said to be gonioisochromatic and exhibiting gonioisochromatism.
These terms are cumbersome and not widely used.
The term metamerism also applies to a pair of light sources
that appear to be the same to one observer but not to another, and that
illuminate one colored object identically but not another object of differ-
ent spectral reflection or transmission characteristics. This phenomenon
is an important consideration in museums and in conservation studios.
The color of a light source is generally described by its correlated color
temperature, K, which does not describe its spectral distribution. Two
light sources may have the same correlated color temperatures but differ-
ent spectral emissions-in essence, exhibiting metamerism. However, the
term metamerism is not applied to light sources. Instead, the degree of
difference of a test light source relative to a reference light source is
described by its color rendering index (CIE 1995). Briefly, as the index
approaches 100 (identity to the test reference lamp), the more closely the
spectral emission of the test lamp matches that of the reference lamp.
Colorimetry 49
Considering the many materials used in producing colors-
colorants, photographic film, the phosphors used in a television tube,
etc.-one realizes the many opportunities for metamerism to occur. It is a
tribute to those individuals who study and work to minimize the magni-
tude of the differences that many observers are not more aware of the
phenomenon.
Other Color Notation Systems
At the beginning of this chapter, it was pointed out that any system for
describing color must be three-dimensional. There can be three bases for
arranging colors in an orderly array: (1) on the basis of appearance (a
perceptual evaluation); (2) on the basis of additive color mixture (psy-
chophysical); or (3) on the basis of subtractive colorant mixture (physical
mixing of colorants). Of the many color systems, only a few will be dis-
cussed here. For more information the reader is referred to the descrip-
tion of methods for describing color given by Judd and Wyszecki (1975);
to the references in appendix E; and to the specific articles cited later.
A book by Agoston (1987) presents the usefulness of color theories and
systems from an art historical point of view, including the theories of
Goethe, Chevreul, and Rood.
The CIE system described earlier in this chapter is based on
additive color mixture, and no collection of color samples exists for this
system. In order to visualize the colors represented by the CIE values, it
is necessary to compare them to real colored materials defined in CIE
notation. One of the most extensive and widely used collections of care-
fully controlled color chips are those representing the Munsell system.
Munsell system
Of the many color-order systems devised, the Munsell system, which is
based on appearance, is most widely used (Munsell 1905; Nickerson
1940). Color chips illustrating the system are readily available in the
Munsell Book of Color. The three dimensions in this system are percep-
tual, and they correspond in general to hue, saturation, and lightness.
Munsell Hue describes hue; Munsell Chroma corresponds to saturation;
and Munsell Value corresponds to lightness. Throughout this book, capi-
talized Hue, Chroma, and Value will be used to identify Munsell dimen-
s10ns.
The system arranges five basic Hues-red, yellow, green, blue,
and purple-in a circle. They are designated by the capital letters R, Y,
G, B, and P. Five intermediate Hues are designated YR (yellow-reds or
oranges), GY (green-yellows), BG (blue-greens), PB (purple-blues), and
RP (red-purples). A numerical prefix is used to further differentiate the
Hues, based on ten steps. For example, starting at the center of the reds,
which is designated SR, and moving clockwise around the Hue Circle,
we find that lOR equals OYR; lOYR equals OY; lOY equals OGY, and so
on. (See the Hue Circle at the bottom of fig. 2.16.) If no number is used
with the Hue designation letter, the assumption is that it is 5, the center
50
Black
01
10PB
5PB
Figure 2.16
Cutaway of the Munsell color solid and the
corresponding constant Hue page, 5Y,
from the Munsell Book of Color (top). The
Hue Circle, with Hue letters and numbers
and Chroma designations radiating from
the center (bottom).
Chapter 2
10BG
5R
10G
5BG
10 / 2 / 4 16 18 / 10 /12 / 14
5YR
6
7
8
Chroma ----->-
9
10YR
5Y
10Y
Colorimetry
of that Hue category. Decimal fractions may be added to the number
prefix to subdivide Hue further, for example, 5.8R.
At the center of the Hue Circle is the gray or neutral point.
51
A perpendicular gray axis through this point, black at the bottom and
white at the top, leads to a three-dimensional array. Ten steps of Munsell
Value (lightness) start at 0 for black and end at 10 for white, with inter-
mediate steps perceived as having uniform differences in Munsell Value.
Again, decimal fractions may be used to define Munsell Value precisely.
The three-dimensional nature of the system is illustrated by
the Munsell color solid in the upper portion of figure 2.16. Radiating
outward from the neutral axis, the colors increase in Munsell Chroma as
the radius increases, so that the purest colors occur at the periphery. The
numerical designation of Munsell Chroma starts at 0 for neutral gray
and increases outward to the most saturated colors, with a Chroma in
the range 14-16 indicating very pure colors. Decimal fractions again
define colors precisely. The Chroma scale is open-ended so that it may
be expanded to include colors approaching the spectral limits.
Munsell devised the following notation scheme for his color-
order system: Hue Value/Chroma, for example, 5.lY 7.8/10. A cutaway
view of the Munsell color solid for Munsell Hue 5.0Y (5.0 yellow) is also
shown in figure 2.16 as a page from the Munsell Book of Color. The chips
in the glossy edition of the book are removable for easy comparison.
The Munsell notations for the three blues discussed earlier
and illustrated in figure 2.10 are
blue 1
blue 2
blue 3
3.73PB 7.66/4.60
0.05PB 7.4 7/4. 79
2.18PB 7.55/4.66
Because the Munsell notation describes visual appearance, it
is always determined from spectrophotometric measurements made with
the specular reflectance excluded. The Munsell notations are included in
accompanying tables for all of the spectral curves in appendix A.
There are many reasons for the popularity and usefulness of
the Munsell system. It is exemplified by real color chips in the Munsell
Book of Color, which is a carefully controlled illustration of the system;
it is an open-ended system in the Chroma dimension so that the develop-
ment of new, vivid colorants and new sources of color can be accommo-
dated; and, last, it has been defined in terms of the 19 31 CIE system for
illuminant C, enabling one to transform between measured CIE values
and Munsell notation. The accepted CIE values are for the renotation
resulting from the smoothing of the system published by Newhall,
Nickerson, and Judd (1943).
It should be emphasized that the Munsell Book of Color is a
representation of the concept of Munsell spacing; it is a carefully made
illustration of the Munsell system. The color, gloss, and spectral curves
of the chips in the book are carefully controlled. Even so, small differ-
ences in chips bearing the same notation do exist for practical reasons
of colorant durability and availability, and the use of new types of more
stable colorants. In addition, there can be small, visible differences in
52
Figure 2.17
CIE 1931 chromaticity diagram showing
loci of constant Munsell Chroma and
constant Munsell Hue at Value 6 (after
ASTM Standard D1535).
Chapter 2
color matching from batch to batch. Thus, the chips in the book are not
primary standards.
In Principles of Color Technology, Billmeyer and Saltzman
(1981) present colored illustrations of a page from the Munsell Book of
Color; a three-dimensional Munsell Color Tree; a diagram of the organi-
zation of the Munsell system; and examples of color tolerance charts.
The relationship between Munsell notation and CIE color
notation is not a simple one. Traditionally, Munsell notation was estab-
lished from measured CIE values by using a table of Munsell Value
versus the tristimulus value Y, and by referral to a published set of chro-
maticity charts (such as shown in fig. 2.1 7) for Munsell Values 1 through
9 at whole-number Value steps, and for near neutrals on an expanded
scale for Values 5 through 9. When the Munsell Value of the sample
determined from the table of CIE Y tristimulus values versus Munsell
Value is not a whole number, as represented in the charts, interpolation
between charts of the two adjacent Values must be made. The table of
Munsell Value versus Y, along with small (about 5 x 6 inches ) conversion
charts and the method for using them, are described in ASTM Standard
D 1535 and in the book by Wyszecki and Stiles (1 982) . The chart for the
level of Munsell Value 6 in CIE 1931 color space is illustrated as figure
2.17. Enlarged charts, 22 x 26 inches, make the task of converting
between systems much easier. They may be purchased from Munsell (see
Munsell Book of Color in references).
This laborious graphical technique was used for many years
to control the Munsell color chips. Today computer programs are avail-
able for calculating Munsell notation (Rheinboldt and Menard 1960).
The reader is warned, however, that the Munsell notations as read out by
0.65
0.60
0.55
0.50
0.45
""
0.40
0.35
0.30
0.25
0.20
0.15
0.10 0.20 0.30 0.40 0.50
x
Hue and Chroma
Value 6
0.60 0.70
Colorimetry 53
some relatively inexpensive measuring instruments may be quite inaccu-
rate. Because of the nature of the CIE-Munsell conversion, one can occa-
sionally get an incorrect notation or no notation at all, even with the best
of programs. In the author's experience, this generally occurs with very
high chroma colors such as can be made with transparent glazes. If the
conversion between CIE notation and Munsell notation needs to be made
frequently, it is a good idea to have a set of the charts, particularly the
enlarged charts, for reference, as well as the Munsell Book of Color for
visual comparisons, in case the computer program fails. With the charts,
extrapolations can be made over a small range for higher chroma colors.
As usual, if no measurement can be made on a particular sample, the
chips in the Munsell Book of Color can be used for a visual estimate of
the color notation by interpolating among the nearest chips in the book.
Special color chips and cards with accurate Munsell nota-
tions-for example, soil charts that describe earth colors-are offered by
Munsell. The company will also make special chips to match a desired
color, as well as tolerance charts representing the desired limits of color
variation about a standard.
Universal color names and dictionary
Another publication that describes a language for color description and
is very useful in a number of ways is Color: Universal Language and
Dictionary of Names by Kelly and Judd (1976). The title of the earlier
edition indicates the origin of the work: The ISCC-NBS Method of
Designating Colors and a Dictionary of Color Names (Kelly and Judd
1955). The ISCC is the Inter-Society Color Council, and the NBS was
the National Bureau of Standards (now known as the National Institute
for Standards and Technology). The original publication was a joint
effort of these two organizations. The 1976 book provides a six-level
system for describing colors in order of increasing preciseness from a
simple color name, such as pink, level 1, up to level 6, with accurately
measured descriptions such as the CIE notation or Munsell notation
calculated from the CIE values. The level just below the samples in the
Munsell Book of Color is the ISCC-NBS Universal Color Language,
level 3. In this level, one of 267 descriptive names is assigned to a color,
based on its Munsell notation, using charts of Munsell Value versus
Munsell Chroma for a range of adjacent Munsell Hues. There are thirty-
one such color-name charts. A representative one is reproduced in figure
2.18 for the Munsell Hue range 9B-5PB, reproduced from the 1976 edi-
tion. Included in this figure are the loci of the three blues discussed ear-
lier in this chapter. In the Universal Color Language they would be
called "pale blue" or "very pale blue," being very close to the line divid-
ing these two regions. This description is its color name in the Universal
Color Language.
Thus, the Universal Color Name can be determined for any
Munsell notation. It may be obtained from comparison with the nearest
Munsell chip in the Munsell Book of Color (level 4) or from visual interpo-
lation between neighboring chips in the Munsell Book of Color (level 5). All
of the pigments described in volume one of Artists' Pigments: A Handbook
54
Figure 2. 18
Page for Munsell Hues 9B- 5PB from Color:
Universal Language and Dictionary of
Names (Kelly and Judd 1976), ill ustrating
the color names assigned for the indicated
Munsell Chroma and Munsell Value blocks,
and the number reference used in t he
Dictionary of Color Names (Kelly and Judd
1955). The three blue colors il lustrated in
figure 2.10 are shown as dots in the blocks
labeled pale blue.
Chapter 2
10/
cu
...-
rn 2
E
O'I $:
IDE 00 ..<::
91
N $:
::i
::0
184
very
>,
pale blue
81
>, bl)
<1l
oii
<:t
\D bl)
m s

Jf ...-
71
..<::
Jf
185
pale
>, blue
61
>,
<1l
bl)

cu


::i
N .::!
<a
-0 ::i
>
cu
::0

5/
E
c
186 ::i

>,
grayish

>, bl)
blue <1l
..<::
4/
N "'
""s

<1l
-0 -""

-0
3/
21
1/
-""
cu
u
::i
<1l ::0
::0
ii
r-..-"" ..<::
\D :;J
V> j
s u
N:O
::0
<1l
::0
rn
00
O'I
00


01
I I I I I
10 /1 /2 /4
1
!. 3
2
183
dark
blue
9B-5PB
180
very
light blue
181
light blue
182
moderate
blue
16 18
Munsell Chroma
179
deep
blue
/10
177
brilliant
blue
178
strong
blue
/12
176
vivid
blue
/14
of Their History and Characteristics (Feller 1986) were described in the
Universal Color Language; Munsell notations were also presented.
Figure 2.18 also shows that each block on the Hue Chart
was reproduced from the 1976 edition and has a number as well as a
name. There are thirty more charts for the other Hue ranges . This num-
ber is used in the "Dictionary of Color Names," which makes up the last
portion of both books. About seventy-five hundred color names are
listed alphabeticall y with references to the source of the names and to
the number of the color-name block in the Munsell Va lue-Munsell
Chroma charts in the 1976 edition. The earlier edition did not have
numbers on the color-name block of the color-name charts. The thirty
sources cited include the Ridgway (1912) collect ion of color plates and
color names, and the Maerz and Paul collection (1930) . Both collections
are of considerable historical interest . Beginning in the last part of the
nineteenth century, more color was used in fine and applied arts (par-
tially, at least, because more colorants became available), and as a result
Colorimetry
more color names were created. The trend continues today. Thus, the
need for the use of the Universal Color Language to describe colors
55
more clearly is still valid. The dictionary portion of the books contains
color names gleaned from past collections of color chips that still exist
and were used in the past. Curators and technical people in art conserva-
tion could identify a color name from this dictionary by which others
could visualize an approximate color. Commercial names could be as
exotic as desired so long as their exotic name was also identified with
the Universal Color Name.
When a need exists to identify a color referred to in the
past, using the color-name dictionary for this purpose may seem diffi-
cult at first. An example may be helpful. If one encountered the name
"Corinthian Pink" and wished to know what the color might have been
that was so described, one could look under the C's in the dictionary for
Corinthian Pink. The following information would be found:
Corinthian Pink
Corinthian Pink
M,m.Pk5; l.gy.R18
R,m.Pk5
The first capital letter gives the reference, and the next series of letters
and numbers describe the color-block area of the color-name charts simi-
lar to figure 2.18. Thus, in the first entry above, M refers to the Maerz
and Paul Dictionary of Color (1930); in the second entry, R refers to the
Ridgway Color Standards and Color Nomenclature (1912). (Copies of
these books may be found in libraries or in antiquarian bookstores. ) The
location abbreviation, m.Pk5, refers to moderate pink in area 5 on the
color-name charts. Since numbering begins on the first color-name chart,
for the Hue range 1R-4R, one looks first on this chart and locates m.Pk5
there. It is important to realize, however, that many of the color names
were used over a range of colors, so it is advisable to check adjacent dia-
grams to see if they are also located there. In this case, m.Pk5 is also
found on a second chart, covering 4R-6R. To determine if the color is
also located on the diagram preceding the first one in the Hue Circle,
the second diagram adjacent to the first one is consulted (the last one,
9RP-1R); this also has an area named m.Pk5. None of the other Hue
Charts contain m.Pk5; hence, the range of Hues where this name was
used corresponds to 9RP-6R. The Munsell Value range of the area is
6.5-8.0, and the Munsell Chroma range is 3.0-7.0. Thus, the range in
the Munsell notation is 9RP-6R 6.5-8/3-7.0. Corinthian Pink is a light,
soft pink, not a strong or vivid pink. Reference to the chips in the
Munsell Book of Color will illustrate the range of colors described.
Maerz and Paul also list l.gy.Rl 8 for Corinthian Pink. That is a light
grayish red in area 18. This area is located on each of the first four
charts, 1R-8R. The Munsell range is expanded to 9RP-8R 5.5-9.012-7.
Hence the color may also have been slightly darker and grayer than the
m.Pk5. The date of the reference to Corinthian Pink would be a helpful
guide in the selection of the probable area to be most representative.
The abbreviations used above are explained in the texts of
both editions. There is also a series of color wheels at Munsell Value lev-
els of 1-8. 7 5. These show the range of the Munsell Hue names used in
56 Chapter 2
the charts. The example of Corinthian Pink was selected deliberatel y to
describe a difficult color, because the m.Pk5 designation occurs on charts
at the beginning and at the end of the sequences. Most color names are
not as broad and are generally described on a single color-name chart.
A great deal of careful research and hard work by a group of
color experts headed by Kell y and Judd went into the preparation of this
dictionary and into the Universal Color Language. The original report,
issued as National Bureau of Standards Circular 553, was published in a
green hardcover in 1955. A later version, with added text by Kell y and
some color plates, was published in 1976. As of this writing, however,
this 158-page booklet is available only as a photocopy. Nonetheless, the
system deserves to be used much more widely.
OSA Uniform Color Scales
Many years of visual evaluations, and the preparation of thousands of
color chips made for the evaluations, culminated in the OSA (Optical
Society of America) Uniform Color Scales System (MacAdam 1974;
Nickerson 1981). Based on equal perceptual differences in all three
dimensions, the OSA system represents the most visually uniform
arrangement created to date. The color chips have been defined in CIE
notation based on the 1964 standard observer and illuminant D65
(MacAdam 1978), and in Munsell renotation (Nickerson 1978). There
are 558 samples in the OSA system, each described in terms of lightness
(L), redness-greenness (g), and yellowness-blueness (j); they are arranged
in a regular rhombohedral-lattice spacing. Each sample, except the edge
samples, is surrounded by twelve others equally visually distant from it.
The system is described by Wyszecki and Stiles (1982) and in ASTM
Standard E 1360. Sets of the chips are currently available from the
Optical Society of America, Washington, D.C.
In the brief section on color harmony presented by Judd and
Wyszecki (1975), their basic theme is that color harmony results from an
orderly plan that is inherently satisfying, a plan that consists of an unam-
biguous selection of colors in the total scene. In the author's opinion, the
OSA system is also an incredibly beautiful work of art. In addition, there
is probably no color-order system that has been so thoroughly docu-
mented in scientific terms. It has been defined in CIE notation based on
the 1964 standard observer and illuminant D65 (MacAdam 1978) and in
the Munsell system (Nickerson 1975, 1978). The preparation of the sam-
ples has been described by Davidson (1978). The history of this extensive
study has been described by Nickerson (1977).
Other systems
Many color-order systems illustrated by physical samples have been cre-
ated in the past, and several new ones have appeared recently. A few of
these are listed here .
The Ostwald color system (Ostwald 1931, 1969) is of histori-
cal interest, even though the atlases of color chips-known as The Color
Harmony Manual, formerly published in several editions by the Container
Corporation of America-are no longer available. The colors of the chips
Colorimetry 57
were based on additive color mixture obtained by disk colorimetry
according to Ostwald's concept that colors can be described in terms of
their pure color combined with white and black. Jacobson (1948) presents
a discussion of the usefulness of the arrangement, particularly to artists
and designers, and provides illustrative color prints. For a recent discus-
sion of the system, the article by Granville (1994) may be consulted.
More recent color-order systems are available and still in use.
One of these is the DIN (Deutsche Industrie Norm) system, the official
German standard color system developed by Richter (1953, 1955). A
description of the system and the CIE notation for the color chips may
be found in the book by Wyszecki and Stiles (1982).
The Natural Color System, which is the Swedish national
standard color-order system, was developed in the 1960s and 1970s
(Hard and Sivik 1981). It is based on the four unique hues-red, green,
blue, and yellow-combined with black and white that are seen in each
color. In this sense it resembles the Ostwald system, except that the
unique hues are imaginary.
The most recent system to appear in the United States is the
Colorcurve system (Stanziola 1992). It consists of an atlas of color chips
with tristimulus values and reflectance factors at 20 nm intervals. These
data and the description of the system are presented in ASTM Standard
E 1541. The system was particularly designed to be an aid in computer
color matching.
Orderly collections of colors are used in the color printing
industry, and they are based on subtractive colorant mixtures of printing
inks. One of these collections is produced by Pantone, Incorporated,
Carlstadt, New Jersey. The problem with producing any printed collec-
tion is that the printing process generates entire pages of colors at once,
so there is no chance to adjust each color to the closest approach to its
aim point. In addition, the reproducibility between successive printings is
difficult to control with high accuracy.
All orderly arrays of color samples and color descriptions are
useful for purposes of communication and visualization. Today the CIE
notation is fundamental in communication, and the Munsell Book of
Color is the most widely used collection of color samples.
Chapter 3
Colorant Characteristics
Generally, materials used to impart color to objects are either dyes or
pigments. Dyes are distinguished from pigments by their solubility in a
solvent or medium; the resulting dye solutions are clear and transparent.
When a dye is incorporated into a compatible solid matrix, such as plas-
tic or glass, the result is a material that may be transparent, like a color
filter, or opaque, if the concentration of dye is sufficiently high to absorb
all of the incident light. When dye solutions are used to color textiles,
paper, etc., the opacity is usually attributable to the base material (fabric,
fibers, paper, etc.), which may be colorless itself but which scatters most
wavelengths of incident light. The dye selectively absorbs some wave-
lengths of the incident light, thus contributing the color, while the sub-
strate scatters the wavelengths of light transmitted by the dye, thus
contributing the opacity. In the case of dyed textiles, the substrate is air,
with a refractive index of 1.00, rather than the textile fibers, including
synthetics, which have indices of refraction in the range 1.46-1.65. This
is because the medium adjacent to the fibers is air, and many dyes are
deposited on the surface of the fibers surrounded by air.
In discussing subtractive colorant mixture, we are concerned
with these two optical characteristics-absorption and scattering of the
incident light . Their relative magnitudes are determined by the chemical
and physical structures of the materials. For a derailed discussion on
the origin of color, the reader is referred to the literature (Nassau 1983;
McLaren 1986; Feynman 1985). The experienced colorist knows that a
number of factors come into play in determining t he behavior of col-
orants, but when all others are equal, it is the light absorption and scat-
tering characteristics of a material that are fundamental.
Absorption of light is a relati vely straightforward phenome-
non. Scattering of light is more complex, and it is encountered with pig-
ments. Unlike dyes, pigments do not dissolve in the system. They exist
as a suspension-dispersion is the term often applied to pigments-that
is, as tiny suspended particles that have the ability to scatter as well as
absorb light. The amount of scattering is a function of the difference in
the refractive index of the pigment and the suspending medium, and of
the particle size relative to the wavelength of the incident light. Reference
to low and high refractive index pigments is relative to that of the typical
transparent binder to which they may be added; the refractive index, n0 ,
of the typical binder is about 1.46-1.65. In this chapter, when we speak
Colorant Characteristics 59
of scattering, we are assuming that, unless otherwise stated, the scatter-
ing is backscattering or multiple scattering, that is, scattering that returns
incident light back toward the light source and to our observing eyes. In
actuality, though, scattering may occur in all directions, both forward
and backward. In some systems, for example, a small addition of Ti0
2
white pigment may result in a perceived darkening of the color, because
most of the incident light is scattered in the forward direction. There
are probably not many situations encountered in the examination of
art objects where this phenomenon is significant. It may, however, be
encountered in glazes and in translucent materials (see chap. 4 ), and
in metallic flake pigmented paints (see chap. 5).
We can divide all conventional pigments (excluding interfer-
ence and metallic flake pigments) into three types: type I pigments scat-
ter much more light than they absorb (for example, white pigments);
type II absorb much more light than they scatter (for example, carbon
black and many organic pigments of small particle size); and type III
both scatter and absorb light significantly (for example, many of the
inorganic pigments).
Type I pigments, which primarily scatter light with little
absorption, are the white pigments. Those with the highest index of
refraction relative to the medium in which they are suspended scatter
the most, for example, the titanium dioxides: rutile at n
0
= 2. 76 and
anatase at n
0
= 2.55. In contrast, lead white (basic lead carbonate
[Pb(OH)
2
2PbC0
3
]) and zinc oxide (ZnO) have refractive indices of
about 2. BaS0
4
(barytes or blanc fixe) has a refractive index of about
1.64, which is close to that of the traditional media in which it is usually
dispersed. The result is that BaS0
4
is not considered a true pigment but
rather an extender or inert (see chap. 6).
Type II pigments, which absorb much more light than they
scatter, include most of the organic pigments as well as the organic car-
bon blacks, lamp blacks, etc. (Black iron oxide, however, is a type III
pigment with a refractive index of about 2.4.) The organic pigments
may be of two types: toners or lakes. Patton (1973) described a toner as
"an organic pigment that is free of inorganic pigment or inorganic carry-
ing base. It is an undiluted organic colorant having maximum tinting
strength for its specific type." He described a lake as "an organic color-
ing matter that has been more or less definitely combined with some
inorganic substrate or carrier."
Type III pigments, which both absorb and scatter light,
include most of the inorganic pigments, such as the iron oxides, and
those organic pigments modified to increase their particle sizes.
Today some synthetic textile fibers are pigmented; the pig-
menting is done in the spinning process. In this case, the pigments exhibit
the same behavior that is encountered in paints or plastics, that is, they
absorb and scatter light. However, there is also significant surface reflec-
tion from the fiber filaments, which adds to the scattering appearance.
(See the discussion on surface reflection in chap. 4.)
Many patterned textiles today are printed with coatings that
are pigmented. As with pigmented printing inks, the colors obtained are
60 Chapter 3
dependent on both the absorption and the scattering characteristics of
the pigments used in the printing formulation. Hiding may not be com-
plete in these applications-a situation similar to that encountered with
glazes used by painters working with both oil and watercolor. (See the
discussion on glazes in chap. 4.)
The degree of both light scattering and absorption is depen-
dent on the particle size of the pigment, as well as its refractive index.
Descriptions of particle size in a given pigment preparation are often
presented graphically in terms of the cumulative percentage of the total
quantity of pigment that exists below each designated particle size. Size
is customarily plotted as the ordinate on a logarithmic scale, and the
cumulative percentage as the abscissa on a linear scale. Such a distribu-
tion is called log-normal. This method of describing particle size is com-
monly used by modern pigment manufacturers and is found throughout
the Pigment Handbook (Patton 1973). A representative graph for Ti0
2
pigments (rutile and anatase) is shown in figure 3.1. The single number
most convenient for describing such a distribution is the median particle
size, that is, the size at the 50% cumulative distribution. In the past, par-
ticle sizes were traditionally determined by their passage through fine
mesh screens of designated particle sizes. Thus, it was the maximum par-
ticle dimension that determined the particle size designation; all smaller
particles passed through the screen.
An accurate method used today for size determination is
counting particle sizes from electron micrographs. Automatic optical
counters have been developed for this use. Electron microscopy also
enables the user to determine the degree of particle dispersion in a poly-
mer matrix. Centrifugal sedimentation is particularly useful for determin-
ing the particle size distribution of very finely divided pigments, such as
many of the organic pigments (Fraser 1973). This technique replaces the
older gravity sedimentation methods described by Orr (1973 ).
For modern pigments, the 50% cumulative distribution size
varies from 0.017 m to 0.5 m for carbon blacks; it is a little larger for
bone blacks (Garrett 1973). Most synthetic organic pigments are gener-
ally less than l. 0 m in diameter (Carr 19 7 3). The median diameter for
Ti0
2
varies from about 0.2 m to 0.25 m (Kampfer 1973). Basic lead
carbonate is about 2.6 m in diameter (in modern production), and basic
lead sulfate is about 1 m (Dunn 1973). The natural iron oxides vary
from 2 m to 5 m (Love 1973) in modern materials, and the synthetic
iron oxides-yellow and red-vary from 0.2 m to 1.0 m (Fuller 1973 ).
The resolution limit for an optical microscope is about 1 m;
below this value particles can be detected only by electron microscopy
(McCrone 1973a). McCrone (1973b) has outlined a scheme for pigment
identification by microscopic techniques.
Modern pigments are manufactured to have particle sizes
providing the desired optical properties according to their refractive
indices. Because the refractive indices of most organic pigments are so
low relative to those of the matrices in which they are suspended, the
particle size of these pigments would have to be increased considerably
to obtain significant scattering, resulting in a high absorption loss. Only
Figure 3.1
Cumulative particle size distribution curves
for typical anatase and rutile Ti0
2
pig-
ments. The graph illustrates the standard
graphing nomenclature used by pigment
manufacturers to describe the particle size
distribution of their product. The abscissa,
a log scale, is the pigment particle diam-
eter, d. The ordinate is the weight percent-
age of all particles below the size indicated
on the abscissa. Thus, the dotted lines on
the graph illustrate that about 26.5 % of
the particles of rutile Ti0
2
are smaller than
0.2 m or, conversely, that about 73.5%
of the particles are larger than 0.2 m. The
size at the 50% cumulative distribution
point is about 0.225 m, as indicated by
the cross (from Kampfer 1973. Copyright
1973 John Wiley & Sons, Inc. Reprinted
by permission).
Colorant Characteristics
0.45
0.40
0.35
0.30
E' 0.25
2-
QJ
N
v;
Q)
Li
t
ro
a..
-0- 0.20
0.15
61
typical ruti le pigment (sulfate process)
typical anatase pigment
0.10 L - ~ - - - - ~ ~ - - - - ~ ~ - - - - ~ ~ - - ~ ~ - - - ~ ~ - - - ~ ~ . . . J . . . . . . ~ ~ . . . J . . . . . . ~ ~ - - - ~ ~ ~
0 20 40 60 80 100
Cumulative undersize particles (wt%)
pigments with refractive indices considerably higher than that of the sus-
pending matrix can scatter light significantly. In the section dealing with
opacity in chapter 4, the theories relating particle size and refractive
index to light absorption and scattering are presented.
There are special types of colorants that may be used to create
particular optical effects in materials. Examples are pearlescent and irides-
cent pigments, and metallic flake or powder pigments, all of which are
discussed in chapter 5, as well as fluorescent pigments or dyes, which
are discussed in chapter 6. However, most colored materials encountered
owe their color characteristics to the absorption and/or scattering of light
rather than to these special effects. Chapter 4 deals with the materials and
objects whose color depends on the absorption and scattering of light.
Chapter 4
Colorant Mixture
Transparent Materials: The Beer-Bouguer Equation
The Beer-Bouguer equation relates the transmittance of radiant energy
through a transparent material (liquid or solid ) to both the material's
thickness and its absorption. It applies to one wavelength at a time.
Sometimes in the literature it is referred to as the Beer- Lambert equation
or simpl y as Beer's law. The Beer- Bouguer equation is a combination of
two earli er ones, which are described below.
Bouguer's (Lambert's) law
An equation relating thickness to transmittance was first formulated by
Bouguer (1729). In 1760 it was "rediscovered" by Lambert (1760).
Although the equation is variously called Bouguer's law or Lambert's
law, the former name has been chosen here because of the earlier exposi-
tion. The law states that the intensity of light passing through an absorb-
ing, nonscattering material is decreased in proportion to the thickness, X,
of the absorbing material with absorption coefficient, A:
log( I/I
0
) =-AX (4. la)
where 1
0
is the intensity of the incident light, and I is the intensity of the
transmitted light . The ratio 111
0
is the transmittance, T. Therefore,
at each A.: -log T = AX (4.lb)
It is assumed that A is a constant so that -log T -or log(J
0
/J)- is pro-
portional to the thickness, X.
A superposing of several layers of materials with different
absorptions and thicknessess is described by t he following form of the
Bouguer equation:
(4.2)
where the A's represent the fixed absorptions of the components identified
with the numbered subscripts, and the X's represent the thicknesses of the
corresponding components. The Bouguer equation is frequently written in
this form. The assumption is made that the multiple layers are in optical
contact, that is, there is no difference in refracti ve index between layers.
The Bouguer equation relating to thickness is customarily called a law
because, as far as it is known, there are no exceptions to it.
Colorant Mixture 63
This is the law used in making complex filters to simulate
the tristimulus functions in tristimulus colorimeters, where the spectral
characteristics of the source used and the spectral sensitivity of the
receiver must also be taken into account. If light illuminates the sample
before passing through the filters, the spectral characteristics of the
light source must be filtered to simulate the desired illuminant. The tri-
stimulus filters must then be made to correspond to each of the color-
matching functions of a CIE standard observer in combination with the
spectral sensitivity of the photodetector, to match as closely as possible
the desired tristimulus spectral responses. Multiple filters are usually
cemented together to achieve this approximation; an adhesive or a
cement with the same refractive index as the filters is used to eliminate
any specular reflection between layers. If the light is filtered to simulate
the tristimulus responses before falling on the sample, a combination
simulating the spectral characteristics of the light source, the color-
matching functions, and the detector responsivity may be assembled into
one so-called filter pack.
There are times when a single layer must be measured in
order to compute its transmittance at a different thickness. In this case,
surface specular reflection must be taken into account because light is
reflected specularly at the surface-air interface where there are differ-
ences in refractive index. The importance of surface reflection in com-
puting colorant formulations is discussed later in this chapter, and the
general role that surface reflection plays in object appearance is discussed
more completely in chapter 6.
When the incident light is perpendicular to the surface, the
light reflected at an interface between two layers having different refrac-
tive indices is calculated by Fresnel's law:
(4.3)
where n
1
is the lower refractive index, n
2
is the higher refractive index,
and k
1
is the decimal fraction of the incident light that is reflected back
at the first interface. (The Greek letter p or simply R may be found in the
literature in place of k to describe the specular or Fresnel reflection at
interfaces.) Unless specified otherwise, the refractive index is measured
at the sodium D lines, about 589-590 nm. If the lower refractive index
medium is air, n
1
is set equal to 1.000. The resulting reflection is called
the specular (mirrorlike) reflection, where the angle of reflection is equal
and opposite to the incident angle. It is also called the Fresnel reflection
because it is based on Fresnel's law (eq. 4.3). Thus if light is incident per-
pendicularly on a piece of opaque polished glass or plastic of refractive
index 1.5, the first surface specular reflectance will be (0.5)
2
/(2.5)
2
=
0.2516.25 = 0.04, or 4% of the incident light.
In the case of transparent materials, such as plastic sheets or
filters, light must also exit when it encounters another air interface. This
results in internal specular or Fresnel reflection-a second reflection due
64 Chapter 4
to an interface. Therefore, the resulting transmittance is decreased fur-
ther as shown in the following:
(4.4)
where p (here used in place of k) describes the interface specular
reflection, T rn is the fractional transmittance measured, and Ti is the
internal transmittance predicted by the Bouguer equation (e q. 4.lb).
In the above example of a perfectly transparent plastic or glass sheet
of refractive index 1.5, with incident light normal to the surface,
and no absorption, the measured transmittance is (0.96)
2
or 0.9216
or 92.1 6%.
All of the above equations describing surface reflectance
calculations are special cases for incident light that is perpendicular to
the surface as described by Fresnel's law. When the angle of incidence
becomes greater than about 10 off the perpendicular, the simple equa-
tion for calculating the specular, or Fresnel, reflection is no longer cor-
rect. The relationship becomes more complex (see eq. 6.2). In this case,
the angle of incidence has to be taken into account as well as the polar-
ization state of the incident light. The Fresnel equations for both hori-
zontally polarized and vertically polarized light are presented by Judd
and Wyszecki (1975). The authors include graphs for the value of the
Fresnel reflection, p, as a function of the angle of incidence for both
the plane polarized light and the perpendicularly polarized light when
the refractive index is 1.5. They point out that for unpolarized light, the
specular (Fresnel) reflection is the average of the two, and they include
this on their graph as well. In a second graph, the authors present the
Fresnel reflection, p, for unpolarized light as a function of the angle of
incidence for materials having refractive indices from 1.1 to 1.9, at
intervals of 0.1.
Beer's equation (the Beer-Bouguer equation)
When thickness, X, is constant, the equation relating the transmission
through a transparent material with absorption coefficient, A, as a func-
tion of the concentration, C, is that of Beer (1852):
or
log(I/1
0
) = -XCA
log(J
0
1I) = XCA
For a mixture of absorbing components, Beer's equation becomes
where X refers to the thickness, the C's refer to the concentration
(4.5a)
(4.5b)
(4.5c)
of absorbing components, and the A's to the specific absorption
coefficients of the components designated numerically. Obviously
Beer's equation is always used with a designated thickness, X, as
described by Bouguer; consequently, it is also called the Beer-Bouguer
equation. In practice, though, the equation is often referred to by the
familiar term "Beer's law, " even though it is not, strictly speaking, a
law, as will be explained later.
Colorant Mixture
For a mixture of materials such as films in optical contact,
where both the thickness and the concentration of each absorber are
variable, the Beer-Bouguer equation becomes
65
(4.6)
It can be seen that changing the concentration has the same effect on the
transmittance as changing the thickness by the same factor. Log(/
0
//)-or
log( l!T)-is the absorbance or the optical density. Many of the spectro-
photometers designed for transmission measurements may also record or
read out in terms of this quantity.
The absorption coefficient, A, is the optical density-log( l!T)
or log(l
0
/J)-per unit concentration and thickness. The coefficient may
be expressed as grams per liter for a 1 cm thickness, sometimes with
the symbol alpha (a); it may also be expressed as the molar extinction
coefficient in moles per liter for a 1 cm thickness, with the symbol
epsilon (E). These are the most commonly used absorption coefficients in
. analytical spectrophotometry. The choice of which to use depends on the
nature of the application. The transparent solvent used in obtaining the
unit absorption coefficient of a compound in solution must be specified
because the absorption spectra may be different in another solvent.
Likewise, the diluent used for solid mixtures must be specified because
the spectral curve may be different with a different diluent (DiBernardo
and Resnick 1959). The concentration units (the C's) as well as the thick-
nesses (the X's) in equations 4.5a-c and 4.6 must be expressed in the
same units as the specific absorption coefficient.
When there is no significant degree of scattering in the mate-
rials, the Beer-Bouguer equation is used in spectrophotometry regardless
of the wavelength of the radiant energy, whether in the visible, the ultra-
violet, or the infrared. When measuring solutions in a solvent placed in
a cell or container of known thickness, it is customary to put a matched
cell containing only solvent in the reference beam of the measuring
instrument (spectrophotometer or colorimeter). When this is done, the
specular reflections from the cell faces cancel out, and no corrections
need be made for interfacial reflections. When measuring solid sheets of
plastic or glass, an undyed or unpigmented sheet of the same thickness
may also be used in the reference beam to cancel our the surface
reflection effects.
Beer's law-the common name often used for the
Beer-Bouguer equation-probably should not be called a "law" because
there are exceptions to its validity; in such cases Beer's law is said to fail.
The equation has been found generally to be valid at low and moderate
concentrations. At higher concentrations, however, many solutions do
not obey Beer's law; the deviations are associated with the formation or
breakup of molecular aggregates and other modifications of the absorb-
ing molecules and the solvent (Wyszecki and Stiles 1982). Stearns (1969)
presents an extensive summary of the factors affecting the results in solu-
tion spectrophotometry; these factors include solvent, pH, concentration,
temperature, interfering ions, oxygen content, and so on. To account for
66 Chapter 4
these factors, it is customary to prepare reference mixtures at a variety of
concentrations to determine the range of concentrations-that is, a cali-
bration curve-over which Beer's law is valid. Often such calibrations
reveal that Beer's law fails if the concentration becomes too high.
Such an observed failure may or may not be due to chemical
interactions. Instead, it may simply be caused by the lack of sensitivity of
the measuring instrument. Remember that the range of transmission mea-
surements is 0-100%. Even if the instrument records optical densities up
to 4, the optical density unit-log(J
0
/I)-is based on the measurement of
the transmittance, I/1
0
. For log(J
0
/I) = 4, the transmittance as measured is
0.01 %. Few instruments are capable of making precise measurements at
such levels. For accurate measurements, then, the concentration must be
decreased by dilution, or the thickness of the material must be decreased.
At the other end of the scale, as transmittance approaches
100%, sensitivity is also limited. Just because the transmittance measure-
ment made on a material is very close to 100%, log(J
0
/I) = 0 does not
necessarily mean that there is no absorption at all. It merely indicates
that the amount of absorption is below the amount that can be detected
by the technique used. If the thickness were increased 100-fold or 1,000-
fold, the same technique might reveal the presence of an absorber.
In the ultraviolet, visi ble, and near-infrared spectral regions,
many compounds absorb because of their electronic structure; often in
the case of organic compounds this is due to conjugated unsaturation
(double bonds ). Spectra of organic materials exhibit extremely high
absorption coefficients: molar extinction coefficients of 10,000 or more,
that is, log(J
0
/I) = 10,000 for a 1 cm thick solution of 1 M (mole per
liter ) concentration. Thus, in order to measure the transmittance of such
a material, it would have to be diluted to a concentration of 0.0001 M,
giving log(J
0
/J) = 1.000, or a transmittance of 10%; this is barely satis-
factory for precision on a good instrument. The range of optical densi-
ties-log(I0/I)-for both visual or instrumental optimum sensitivity is
0.25 to 0.8, corresponding to transmittances from 56% down to 16%. In
the example of an absorber with a molar extinction coefficient of 10,000,
a concentration of 0.0001 M for an absorber with a molecular weight of
250 would correspond to a weight concentration of 0.025 g 1-
1
Thus, if
one were to have a sample containing this material dissolved in 10 ml of
solvent, the presence of 0.00025 g of the material could be detected at
log(J
0
/J) = 1.000; similarly, 0.00010 g of material could be detected if
log(J
0
/J) = 0.4 instead of 1, a very good range of concentration.
It is not unusual for molar extinction coefficients of dyes and
pigments in solution to be many times higher, even ten times or more,
than the example above. Thus very small amounts of colorants may be
detected, and they may have a significant effect on perceived color (see
"Comments on Colorant Calculations and Identifications" later in this
chapter). This great sensitivity is one of the chief advantages of the solu-
tion spectrophotometric technique for determining very small amounts of
colorants, dyes, or pigments. Today Saltzman is the most knowledgeable
proponent and exponent of this technique (Saltzman and Keay 1967;
Saltzman 1986). Kumar (1981) describes in helpful detail a series of ana-
Colorant Mixture 67
lytical schemes for identifying organic pigments this way. Kumar's
research, partially funded by a grant from the National Museum Act,
included curves of more than 150 different colorants and illustrated the
application of the technique to paints, plastics, printing inks, and many
artists' materials. (Reference curves are available from Billmeyer. See ref-
erences section for address.) Although the solution spectrophotometric
technique requires taking a very small sample of the object, it is useful
where necessary.
Another microtechnique useful in analyzing very small
samples is thin-layer chromatography coupled with microspectro-
photometry. Fuller (1985) describes this technique and presents several
examples of its application in forensic science.
Another useful technique, described more completely in chap-
ter 7, is that of dispersing a small sample of a dark or high-chroma color
in a white-pigmented base using a micro ball mill, painting an area, mea-
suring it when dry, and examining the reflectance curve.
It should be kept in mind, however, that reflectance spectro-
photometry is a useful first step in analyzing materials because it is
nondestructive.
The discussion in this section applies primarily to organic
materials, including dyes and dissolved pigments. Generally speaking,
inorganic materials in solution do not absorb as highly or as specifically
as organic materials. Libraries of curves of the electronic absorption
spectra of compounds in solution as measured in the ultraviolet or the
visible regions are listed by Stearns (1969).
Absorbing and Scattering Materials: The Kubeika-Munk Equation
The most widely used equation relating the reflectance to the concentra-
tion of pigments is that of Kubeika and Munk (1931) and of Kubeika
(1948, 1954). The pigments are characterized in terms of their absorp-
tion, K, and scattering, S. When hiding is complete (no substrate color is
visible), the Kubeika-Munk equation states that at each wavelength (A),
[1 - R(A.)]212R(A.) = K(A. )IS(A.)
(1-R)
2
K
or simply ---= - (at each wavelength, A.)
2R S
(4.7)
where R is the decimal fractional reflectance at wavelength A.; K(A. ) is the
total absorption in the material at the same wavelength; and S(A. ) is the
total scattering in the material at this wavelength. (Generally, the desig-
nation (A) is omitted for simplicity.) The Kubeika-Munk absorption
coefficient K is not the same as the absorption coefficient A used in
the Beer-Bouguer equation for transmittance (eq. 4.6). The Kin the
Kubeika-Munk equation (eq. 4.7) is twice the A in equation 4.6 when S
approaches zero (Judd and Wyszecki 1975) . In addition to the limitation
that it applies to one wavelength at a time, the simple Kubeika-Munk
relationship (eq. 4.7) applies only to materials that are completely
68
Figure 4.1
The relationship between the percent
reflectance and the ratio K!S calculated
from (1-R)
2
/2R = KIS (eq. 4.7), in which
R is expressed as a decimal fraction.
Chapter 4
opaque, that is, for those materials that transmit no light through to
the background. (When a pigmented material is not completely opaque,
other forms of the equation must be used that take into account the
color of the background or the substrate. See the sections on glazes and
opacity later in this chapter.) Note that the ratio K/S is dimensionless; K
and S must be expressed in the same units based on weight or volume.
This simple relationship tells us that if we increase the total
absorption in the material without changing the total scattering, the
reflectance will decrease. Conversely, if we increase the total scattering
without changing the total absorption, the reflectance will increase.
Figure 4.1 shows a graph of the relationship between KIS and the
reflectance, R, as defined in equation 4. 7.
The Kubeika-Munk equation states mathematically the
well-known fact that if we add black (which has a high Kand absorbs
strongly), we lower R and darken the color; if we add white (which scat-
ters strongly), we increase R and lighten the color. A half century ago,
Saunderson (1 942) showed how the equation could be used in calculating
colorant concentration in some plastics.
20 40 60 80 100
Percent reflectance
Colorant Mixture 69
Duncan (1962), working with paints, showed that the total
absorption of a system made with a mixture of pigments is the sum of
the fractional amounts of absorption from each pigment based on the rel-
ative concentration, and that the total scattering is also the sum of the
fractional amounts of scattering from each pigment present. The basic
Kubeika-Munk equation then becomes
(l-R
2
) =(KJ = C1K1 +C2K2 +C3K3 + .. .
2R s M C1S1+C2S2 +C3S3+ .. .
(4.8)
Note that the designation (A.), meaning "at each wavelength," has been
omitted for simplicity, as described earlier. Here the subscript M indi-
cates the mixture, and the subscript numerals identify the pigments in
the mixture. The C's are the relative fractional concentrations of the
individual components in the mixture, such that the total of the concen-
trations is unity. Again, the assumption is made that hiding is complete-
the layer is opaque. The K's and the S's must be in the same units of
concentration. It is this form of the equation that we consider first, the
form that is widely used for color matching with the aid of a computer.
Generally, this equation is solved simultaneously at a minimum of 16
wavelengths. In chapters 4 and 5 of Computer Colorant Formulation,
Kuehni (1975) presents a summary of the techniques used in writing a
computer program for colorant matching using equation 4.8.
One of the factors that the above relationship (eq. 4.8) tells
us, and which needs to be emphasized, is that when the colored material
is completely opaque, the relative concentrations of pigments in the mix-
ture determine the reflectance; the color is, therefore, not dependent on
the absolute concentrations of pigments present. This is not strictly true,
as the trained colorist knows, but the color differences encountered when
changing the absolute pigment concentrations are small unless the critical
pigment volume concentration (CPVC) is exceeded (see chap. 6), result-
ing in changes in gloss and pigment efficiency.
When used alone with no other pigments present, that is, as a
masstone, type II pigments described in chapter 3 (pigments that absorb
strongly relative to their scattering) look almost black and cannot be
identified from their spectral reflectance curves. Examples are carbon
black, phthalocyanine blue, carbazole dioxazine violet, phthalocyanine
green, and some quinacridone reds and violets-all are transparent
organic pigments. This fact explains why some mixtures of white and a
chromatic pigment "loop around" on the chromaticity diagram as their
relative concentration in mixture with white is increased. Since they
approach neutral in very high concentrations, there is a critical concentra-
tion in admixtures with a specified white where maximum purity or satu-
ration occurs, as illustrated in figure 2.12 (Johnston 1973). Examination
of the loci shows that the pigments that do not exhibit near-straight-line
progressions on the chromaticity diagram are all type II organic pigments,
those that absorb strongly relative to their scattering.
Because their masstone color is nearly black, they are not
used singly but are either mixed with a scattering pigment, such as white,
70 Chapter 4
or applied at incomplete hiding over a reflecting substrate (see sections
on glazes and opacity later in this chapter). In both types of application,
the reflectance spectrum in the absorption region shows the individual
absorption bands that are useful for identification.
Those pigments that exhibit significant reflectance when
used in a masstone are either type I nonabsorbing white pigments or
type III absorbing and scattering pigments. When mixed with decreasing
amounts of white, type III pigments continue to move in a relatively
straight line on the chromaticity diagram all the way to the masstone
color, as illustrated by molybdate orange and light chrome yellow in
figure 2.12. Both are inorganic, absorbing and scattering pigments. As
the scattering from the lessening amount of white decreases, the loss is
compensated by the increased scattering due to the increasing concentra-
tion of the chromatic pigment.
The organic yellow FGL shown in the figure illustrates that
not all organic colorants exhibit low scattering. Factors that affect the
absorption and scattering strengths of pigments are discussed later in
this chapter.
Qualitative Application of the Kubeika-Munk Relationship
The reflectance curves in appendix A will be referenced for the initial dis-
cussion of the techniques for qualitative identification of colorants by
spectral reflectance curve analysis. These curves comprise four white pig-
ments, carbon black, and eleven chromatic colorants. Appendix A also
includes tables of the CIE and Munsell notations based on nonspecular
curves (total percent reflectance minus 4 % ) .
The achromatic colorants-the whites and the blacks-have
few characteristic clues for identification by reflectance spectrophotome-
try because they either absorb all the way across the spectrum, as does
carbon black (fig. A.3), or essentially scatter throughout, as do the
whites. The one white pigment exception is rutile Ti0
2
Figure A.3 shows
the reflectance curves of carbon black in mixture with rutile Ti0
2
at sev-
eral concentrations. At all concentrations shown- from 0.1 % black and
99.9% Ti0
2
, to 8% black and 92 % Ti0
2
-the reflectance decreases as
the wavelength decreases below 420 nm. At 0% black (100% Ti0
2
) , the
same increase in absorption (decrease in reflectance) occurs. This absorp-
tion is characteristic of rutile Ti0
2
Because absorbances in mixtures are
additive, this characteristic frequently appears; when seen in curves of
mixtures with white, it indicates rutile Ti0
2
The carbon black has no
absorption maximum. The maximum reflectance occurring at 420 nm in
figure A.3 is caused by the scattering maximum of the rutile Ti0
2
in this
region. Anatase Ti0
2
white begins to absorb only below about 405 nm.
If this same carbon black is mixed with different white pig-
ments, such as anatase Ti0
2
, zinc oxide (ZnO), or lead white (basic lead
carbonate [Pb(OH)i2PbC0
3
]), the spectral curves are different because
the coefficients of absorption and scattering of the white in the mixtures
are different. Figure A.4 shows the spectral curves of four mixtures of
Table 4.1
Tristimulus values and chromaticity coordi-
nates for the white plus trace of black
mixtures in fig. A.4 (see appendix A) (CIE
illuminant C, 1931 standard observer).
Colorant Mixture 71
carbon black with these white pigments. Note that the concentration of
black is different for each of the mixtures, taking into account differ-
ences in tinting strength of the whites. (The tinting strength of white pig-
ments depends on their scattering strength; this is discussed later in this
chapter.) The differences in concentration were chosen to give spectral
curves of similar luminous reflectance (tristimulus value Y ), so that com-
parisons of the spectral reflectance curves and the resultant color of the
mixtures could easily be made.
Table 4.1 gives the tristimulus values and chromaticity coor-
dinates for the four mixtures based on diffuse reflectance measurements.
Figure A.2 shows the data in table 4.1 plotted on a chro-
maticity diagram. Although all of the mixtures are bluish, the lead white
mixture is most nearly neutral (closest to the illuminant), corresponding
to the greater flatness of the spectral curve. These mixtures of black and
white pigments demonstrate the scattering characteristics of the whites.
To picture the absorption characteristics of the whites, a curve of each
white pigment by itself (masstone), based on a complete hiding film, can
be examined in figure A.5. From the curves it can be seen that rutile
Ti02 absorbs most strongly in the violet, and that the other white
pigments absorb much less. The corresponding Munsell notations are
given in table A. l.
It must be emphasized that these data are based on specific
samples of the pigments, but some general statements can be made: (1)
Rutile Ti02 always absorbs in the violet, giving it a yellowish hue in
masstone. During manufacturing, the particle size of many rutile Ti0
2
pigments is adjusted to have a scattering maximum in the violet region
in order to offset, to some extent, the yellowness due to its violet absorp-
tion. (2) It is difficult to distinguish, on the basis of the spectrophoto-
metric curves alone, among anatase Ti0
2
, ZnO, and lead white when
mixed with chromatic pigments.
White pigments, being of mineral nature, can be identified by
X-ray fluorescence or, if a tiny sample can be taken, by X-ray diffraction.
Levison (1976) points out that Ti0
2
(anatase in oils, primarily, and rutile
in water-based paints ) is often diluted with blanc fixe (BaS0
4
) or with
ZnO for better working properties in artists' paints (see chap. 6). He also
reports that the addition of a small amount of ZnO to the Ti0
2
retards
the yellowing of linseed oil or safflower oil vehicles.
Though not a neutral in the sense of not exhibiting any
significant wavelength selectivity, burnt umber is a favorite near neutral
to replace black on artists' palettes, even though it is yellowish. When
industrial tinting (or shading) was done by visual means alone, burnt
umber was also a favorite of tinters (or "shaders") because it is easier to
Mixture x
y
z x y
1 % black+ R. Ti0
2
37.59 38.65 51.32 0.2947 0.3030
0.95% black+ A. Ti0
2
36.08 36.98 52.18 0.2881 0.2953
0.1927 % black + ZnO 38.90 39.57 57.87 0.2853 0.2902
0.2275% black+ Pb white 40.45 41.34 52.34 0.3016 0.3082
72 Chapter 4
control its effect on mixtures than it is to control the effect of the more
strongly absorbing carbon blacks. Since burnt umber is a natural iron
oxide pigment, its permanence is very good.
Figure A.6 shows the spectrophotometric curves of several
concentrations of burnt umber in mixture with rutile Ti0
2
It can be seen
that the resultant colors are yellowish compared with the more neutral
carbon blacks; burnt umber has selective absorption in the violet and
blue regions up to about 530 nm. Even in masstone, it exhibits a dark
yellowish color. The reason for this persistent yellowness is simply that it
also scatters significantly in the yellow-orange and red regions. We shall
see later how important thi s characteristic is.
Because of the wide use of burnt umber, all of the chromatic
pigments in the abbreviated library in appendix A were compared in
mixture with rutile Ti0
2
and with burnt umber, as well as with black.
Table A. l gives the CIE and Munsell notations for both the black-Ti0
2
mixtures and for the burnt umber-Ti0
2
mixtures. The reflectance curves
for one mixture of each colorant in rutile Ti0
2
, alone and with black and
with burnt umber, are illustrated after the corresponding series of mix-
tures made with rutile Ti0
2
.
The primary chromatic pigments are blue, red, and yellow;
they will be discussed next.
The reflectance curves for ultramarine blue (fig. A.8 ) and
for phthalocyanine (phthalo) blue (fig. A.10) show that the maximum
absorption (minimum reflectance) occurs in the complementary wave-
length (orange) region at about 600 nm. The shape of the absorption tail
on the phthalo blue at 600-700 nm is highly characteristic and provides
a clue to the pigment's presence. The ultramarine blue is less distinctive,
having a broad region of absorption that decreases in the red region with
a characteristic red reflection tail in its spectrum. This gives it a more
reddish hue than that of phthalo blue, as can be seen from the Munsell
notations in table A.2.
Figure A.11 shows the effect on the curve shape of adding
a small amount of black to the phthalo blue-rutile Ti0
2
mixture illus-
trated in figure A.10. It is apparent that the addition lowers the reflec-
tance in the blue (maximum reflectance) region significantly but does
not change the reflectance in the red (minimum reflectance, maximum
absorption) region nearly as much. Because the absorption, K, is very
low in the blue region (high reflectance, so that it looks blue), the addi-
tion of the strongly absorbing black makes a significant change in the
K/S ratio in the blue region; this lowers the reflectance, R, drastically. In
the orange region, however, the addition of black has far less effect on
the reflectance because the absorption, K, of the phthalo blue is already
very high in this region.
With the addition of burnt umber, the absorption of the
umber in the violet becomes apparent in the curve, leading to yellowness
by shifting the reflectance maximum significantly into the green region.
Note, however, that because of the long-wavelength scattering of the
burnt umber in the orange and red regions, the added absorption in this
range is partially offset by the moderately high scattering of the umber.
Colorant Mixture 73
The CIE and Munsell notations for the blue curves are presented in table
A.2. Figure A. 7 shows a chromaticity plot of the colorants mixed with
white, and with black or umber added. The yellowing effect of the burnt
umber is apparent.
The mixtures of ultramarine blue with rutile Ti0
2
(fig. A.8),
and also those mixtures with small amounts of black or umber added
(fig. A.9), further illustrate the principle of how the sum of absorption
and scattering in a mixture affects the reflectance. Figure A.9 illustrates
how the addition of two concentrations of black affect the curve shape.
The reflectance maxima for both of the mixtures with black are in a
bluer region and are higher than the maximum for the mixture made
with burnt umber, even though the curves in the absorption region for
two of them-0.02 % black and 0.5 % burnt umber-are similar. The
short-wavelength violet absorption of the burnt umber dominates the
resultant hue, making it more greenish. Notice, however, that the long-
wavelength scattering of the umber raises up the reflectance to nearly
that of the lowest amount of black, 0.02 %.
Thus, it can be seen that, on a qualitative basis of interpreta-
tion, the relationship of the resultant reflectance to the ratio of the total
absorption and scattering makes it possible to understand much about
pigment mixtures: the characteristic absorptions are not obscured in mix-
tures. When pigments scatter, that effect must also be taken into account,
as in the case of the umber. It should be pointed out, however, that pig-
ment scattering is generally not highly spectrally selective but exhibits
relatively gradual changes with wavelength. Note also that the absorp-
tion bands of pigments in mixtures, when the bands occur in the scatter-
ing region of another pigment, are not obscured even in the region of
significant scattering from the other pigment in the mixture.
Figure A.12 shows the spectral reflectance curves of red iron
oxide in mixture with rutile Ti0
2
This particular sample of red oxide is
borderline to an orange. However, all red iron oxides have similar curves
absorbing in the violet, blue, and green regions and reflecting the long
wavelengths. The wavelength at which the reflectance rises determines
the hue; the longer the wavelength at which the rise occurs, the redder, or
less orange, is the pigment. This particular sample is of synthetic origin,
but the natural red oxides exhibit similar absorptions, although they are
generally lower in chroma, reflecting less in the red region. There are no
distinguishing features in the absorption region other than the flatness.
However, this pigment exhibits a characteristic reflectance in the red
region, rising continuously after 620 nm in a characteristic upward pat-
tern. Both natural and synthetic red iron oxides exhibit this feature.
Bright reds with little or no white (masstones or near mass-
tones) exhibit flat, low-reflectance curves in the region below about 600
nm. (See fig. A.14 for masstone BON red, 100.0%; this is an organic red.)
Therefore, as mentioned earlier, the spectral reflectances of masstones aid
little in identification. However, when a bright red is mixed with a scatter-
ing white pigment, one may often observe characteristic absorption bands.
Some, like the BON red illustrated, exhibit a single strong absorption
maximum in the white mixture. Others, such as Quinacridone Red
74 Chapter 4
(fig. A.16), have two or more distinctive absorption bands. Their curve
shape is specific to Quinacridone Red; no other red pigment has this
unique fingerprint. Even when the red is mixed with other pigments, the
pattern remains intact. Figures A.15 and A.17 illustrate the effect of
adding black or burnt umber to the BON red and to the Quinacridone
Red, respectively. For both red pigments, the addition of black lowers the
reflectance significantly in the long-wavelength, high-reflectance region.
The burnt umber also affects the short-wavelength region, shifting the hue
a little toward orange. All of the visual, perceived changes are reflected in
the CIE and Munsell notations in table A.3. Figure A. 7 shows the hues
and dominant wavelengths on the chromaticity diagram.
The remaining primary subtractive color to be considered is
yellow. Yellow pigments absorb in the short-wavelength, blue-violet
region of the spectrum-up to 480 nm, for example. Most yellows
exhibit only one absorption band in the visible region of the spectrum.
Figure A.18 for yellow iron oxide and figure A.20 for Permanent Yellow
FGL are illustrative of yellows in mixture with rutile Ti0
2
There are
subtle differences we can sometimes use as a possible clue to identify
each yellow, but it is difficult to be absolutely certain solely on the basis
of the reflectance in the visible region.
The mixtures of the yellows with black and umber illustrate a
well-known problem in graying colors with black. Figures A.19 and A.21
illustrate the reflectance curves of such mixtures. Mixing carbon-black
pigments, which absorb and do not scatter light, with yellow pigments,
such as the yellow oxide and the FGL, that do scatter, results in a green
color because yellows exhibit their maximum scattering in the green
region. Indeed, olive greens are frequently made this way-by the admix-
ture of yellow and black; they contain no green pigment at all. Graying
with burnt umber instead of black alleviates this tendency to shift toward
green. In table A.4, the Munsell notations reflect this difference. Refer to
the Munsell Hue Circle in figure 2.16. With just white and 10% yellow
oxide, the Munsell Hue is 1.51 Y. With black added, the hue moves
toward the green, to 9 .09Y. With umber added, the hue shifts toward
the red, to 9.83YR. The Permanent Yellow FGL at 1.00% in Ti0
2
has
hue of 9. 71 Y. With black added, it shifts to greenish yellow, 4.24GY;
with umber, it shifts back to more reddish yellow, 4. 79Y. These changes
are very apparent on the chromaticity diagram (fig. A. 7).
The secondary colorants are the greens, the purples (often
called violets), and the oranges. They are called secondary because the
hue can be made by the admixture of two of the primary pigments: green
by mixing blue and yellow pigments; purple by mixing red and blue
pigments; and orange by mixing red and yellow pigments. Secondary
colors, however, can also be distinctive pigments and not mixtures. When
observing a colored material visually, therefore, we cannot be certain
whether a secondary color is based on a single pigment in mixture with
white, or if it is an admixture of two primary colors.
The only green pigment illustrated in the limited library given
in appendix A is phthalocyanine green (phthalo green) (see fig. A.22).
Note the characteristical ly shaped tail in the red region, 620- 700 nm.
Colorant Mixture
This shape uniquely identifies phthalo green. Note that its absorption
maximum is in the red at about 650 nm.
75
Reflectance curves of mixtures of phthalo green in Ti0
2
white
with black or burnt umber are illustrated in figure A.23. The addition of
both of these pigments decreases the reflectance in the green, but the burnt
umber moves the reflectance maximum to longer wavelengths so that it
appears more yellow. CIE and Munsell notations are given in table A.5.
When examining unknown green pigments, the reflectance
region is not the characteristic region. It is the minimum reflectance
(maximum absorption) region that is characteristic. When using spectral
reflectance curves for identification, the sleuth should begin examining
the curve at the long-wavelength red region where greens and blues
absorb. The reason for this is that reds, oranges, and yellows do not
absorb significantly in this long-wavelength red region, and neither do
the purples, but most green and blue pigments do.
Compare the shape of the green made with a single pigment,
phthalo green (see fig. A.22), with those of the greens made by admixture
of blues and yellows illustrated in figure A.24. Curves 1 and 2 in figure
A.24 have obviously been made with a phthalo blue: the fingerprint long-
wavelength tail uniquely identifies phthalo blue. The differences between
the two curves are due to the use of the two different yellows. Curve 1 is
highest in chroma; it is made with the higher-chroma yellow among our
reference pigments-Permanent Yellow FGL (fig. A.20). The reflectance
in the blue region rises at a shorter wavelength than does the reflectance
of the yellow oxide in our limited library. See figure A.18. Given a choice
between these two yellows, the selection of the yellow that was used is
easy. However, in an unknown palette, the identification is far from cer-
tain. The presence of the phthalo blue, though, is rather certain. This
example points out a very important aspect of the reflectance curves of
yellow pigments: the wavelength at which the reflectance rises at the
short wavelength limits the selection of possible yellows used.
The other two greens in figure A.24 are obviously made
with the other blue, ultramarine blue, in the limited library of examples
(fig. A.8). Again, choosing between the two yellows in our library, it is
easy to see that the mixture of curve 3 is made with the higher-chroma
Permanent Yellow FGL, and that of curve 4 is made with the lower-
chroma, redder, yellow iron oxide. CIE and Munsell notations are given
in table A.5.
Purples reflect in both the red and the violet regions, being
complementary to green or yellow-green, the spectral region in which the
purples have their absorption maxima. In our collection of pigments in
appendix A, there is only one purple: carbazole dioxazine violet, which
will be referred to as simply carbazole violet. It is illustrated in figure
A.25. Note that this purple exhibits three characteristic absorption
bands, giving a very distinctive curve shape in the absorption region. In
mixtures with black and umber (fig. A.26), this shape is still apparent.
Contrast this shape with that for the violets in figure A.27
made by the mixture of blues and reds. Beginning in the long-wavelength
region, we immediately identify phthalo blue in curve 3. The red used
76 Chapter 4
with it is also unmistakable: the three absorption bands uniquely identify
Quinacridone Red. Examining curves 1 and 2 in the long-wavelength
region, we see that they are made with ultramarine blue, the other blue
in our limited library in appendix A. If we had many more blue pigments
to contend with, our identification would not be quite as certain, but
we could say with certainty that phthalo blue was not used in a major
amount if its characteristic absorption is not present. In these two curves
we cannot identify the reds with certainty, except within the limitation of
our library. Curve 1 is higher in chroma than curve 2, indicating it may
have been made with a red higher in chroma. Our only other organic,
high-chroma red is the BON red (fig. A.14 ). The absorption shape, with
a maximum at 5 60 nm, suggests that this could be the red used. The
mixture of curve 2 could have been made with red oxide because of its
relative flatness in the violet and blue regions. We can say with certainty,
however, that Quinacridone Red was not used to obtain curves 1 and 2;
its characteristic three-band absorption shape is not present. CIE and
Munsell notations are presented in table A.6, and chromaticity graphs
are shown in figures A.1 and A. 7.
Our final secondary color is orange. Oranges are comple-
mentary to blue. In our limited library there is only one orange-molyb-
date orange (fig. A.28). This particular inorganic orange absorbs in the
violet, blue, and green regions, making a very flat reflection curve in this
region, with no distinguishing bands. In mixture with black and umber,
this orange (fig. A.29) produces curves similar to those in figure A.13 for
red iron oxide. However, note that the upward tail of the red oxide is
a clue to red oxide-although a subtle one-to distinguish it from the
molybdate orange.
Oranges can be made by mixture of red and yellow. Some
mixtures made with the pigments in our library are shown in figure
A.30. Beginning our examination at the long wavelength, the reflec-
tance tail on curve 3 suggests that red oxide (fig. A.12 ) is present.
Neither curve 2 nor curve 1 has any character in this region. Working
down to the middle of the spectrum, curve 2 should catch our eye. The
shape is reminiscent of Quinacridone Red, except that the absorption
band of the shortest wavelength of the three characteristic bands is not
readily visible; it has been obscured by the strong absorption of a yel-
low pigment. Because of the high chroma we chose Permanent Yellow
FGL as the strong absorber in this mixture. Curve 1 is also a high-
chroma orange. The absorption maximum at 560 nm suggests that
BON red is present. Given an extensive choice of reds, however, no
such identification could be made because many high-chroma reds have
only a single absorption band (see appendix D). The shape of the short-
wavelength absorption, which begins to decrease at 4 70 nm, suggests a
high-chroma yellow pigment. From our library we select Permanent
Yellow FGL. We could not identify the yellow on the basis of spectral
curve shape alone, however. We could say it is not yellow oxide
because that pigment does not begin to decrease in absorption until
about 500-510 nm. For comparison, the curve for molybdate orange
(curve 4) is included in figure A.30. Munsell and CIE notations are pre-
Colorant Mixture 77
sented in table A. 7, and chromaticity diagrams are illustrated in figures
A.1 and A.7.
Probably the most difficult colors to match precisely, with no
significant degree of metamerism, are the near neutrals. Grays and tans
can be made using a wide variety of mixtures. Two examples of such
chromatic pigment mixtures are illustrated in figure A.31. Included for
comparison are the flat curves for 0.5% and 2.0% black. In examining
the three chromatic mixtures, we again begin in the long-wavelength
regions. The fingerprint for curve 1 reads phthalo blue and for curve 2,
phthalo green; refer again to figures A.10 (phthalo blue) and A.22
(phthalo green). The next lower-wavelength absorption maximum on
curve 2 is at 560 nm. Checking our reference curves in appendix A we
find that BON red (fig. A.14) has a maximum absorption at this wave-
length. Curve 3 is much less characteristic, but we would expect from its
flatness that there could be an orange or orange-red to absorb in the vio-
let, blue, and green regions. In this case red oxide was used. Curve 3 is
not very characteristic, but it has an indication of an absorption band at
600 nm, in the orange region, so it must contain a blue. It is obviously
not phthalo blue as used in curve 1, but it certainly could be ultramarine
blue because of the red upward tail (see fig. A.8 ). The next slight indica-
tion of an absorption band is at about 550 nm. This could be caused by
an orange such as molybdate orange (fig. A.28) or an iron oxide orange
made with red oxide (fig. A.12) and yellow oxide (fig. A.18), for example.
As indicated in figure A.31, the latter is the way it was actually made.
In this section we have illustrated the basic principles used in
the qualitative analysis of spectrophotometric reflectance curves, utilizing
a very limited library of reference pigments. In essence, the approach to
analyzing the shape of the reflectance curve is to start the examination for
characteristic absorption (low-reflection) bands in the long-wavelength
region, and to proceed in the examination progressively toward the short-
wavelength region. We have also tried to show how the reflectance curves
relate to our visual observations, and how they relate to the derived col-
orimetric descriptions, to the CIE coordinates, and to the Munsell nota-
tions. The usefulness of plotting the colorimetric coordinates has been
emphasized. In short, the major point of this section is that we should use
every tool derived from the measured spectrophotometric curve in our
analysis-colorimetry as well as spectrophotometric curve shape.
The limited number of pigments used for illustration repre-
sents only a small fraction of the pigment types that might be utilized by
an artist. Later in this monograph we describe and discuss other pigments
and their characteristic reflectance curves, many of which are illustrated
in chapter 7. It is hoped that a library of reflection curves can be accumu-
lated for future reference, similar to the atlas of infrared curves published
by the Federation of Societies for Coatings Technology (1991) and to the
solution spectra sources summarized by Stearns (1969). An example of
how such a library could be organized is presented in appendix D.
This limited number of pigments-chromatic colorants plus
black, white, and umber-is probably about the maximum number of dif-
ferent pigments an artist might use in any single painting. By measuring
78 Chapter 4
many areas of color on a painting, one can get an idea of the palette the
artist may have used or at least narrow down the possibilities. Chapter 8
describes a protocol for organizing many multiple measurements and eval-
uations. If all we can say for certain about many areas is that they were
not made with certain pigments, we have, at least, eliminated some of the
possibilities. We may confirm the presence of the inorganic pigments by
X-ray fluorescence. A consideration of the information obtained by these
nondestructive techniques allows the examiner to limit the number of
samples that need to be taken from a valuable museum object .
Quantitative Application of the Kubeika-Munk Equation
The Kubeika-Munk equation described earlier in this chapter (eq. 4.8 )
describes the separate additivity of the K's (a bsorption) and S's (scatter-
ing) for the pigments in a mixture. Each of the K's and S's is multiplied
by the relative concentration of the pigment it represents. Since the K's
and S's refer to the absorption and scattering of unit concentrations of
the pigments, they are called unit K's and S's. They may be determined
from reflectances measured on samples of known concentration for
selected pigment mixtures and on masstones.
Simple estimation
In the previous section, the qualitative application of equation 4.8 was
based on measurements of opaque mixtures of individual pigments with
white. In this instance, most of the scattering comes from the white pig-
ment. What is observed in the characterization of each chromatic pigment
in this manner is its characteristic absorption, K, as a function of wave-
length, relative to the major scattering from the white pigment, Sw. From
this point of view, equation 4.8 can be rewritten in the following form:
where the subscript numbers refer to the individual pigments in the mix-
ture, M. This equation applies at each wavelength, but for simplicity we
have omitted the designation (A) that should always follow R, K, and S
to indicate their wavelength dependency.
Since the scattering of the white pigment, Sw, is constant in
the equation, and it is only the rati o of K to S that matters for opaque
samples, Sw can be set equal to 1.0000 at all wavelengths. Now the unit
Kp/Sw for each pigment can be calculated from a mixture of each chro-
matic pigment, subscript P, with the white pigment, subscript W, by
arranging equation 4.9 for one pigment, P, in mixture with white. Note
that Sw must be that of the same white pigment in the same vehicle for
all of the unit (Kp/Sw) values used.
This equation works well for mixtures where most of the
scattering is due to one white pigment. It is also used for dyes on fabrics
that supply the scattering.
Colorant Mixture 79
Since only the ratio K/S is used, these calculations are known
as single-constant Kubeika-Munk calculations. From the measured
reflectance, the KIS ratio can easily be obtained using a small desk cal-
culator. However, there are published tables of the value of KIS for any
reflectance greater than 0.1 %, up to 100.0% at 0.1 % intervals; one such
table is reproduced in appendix B. Recall that the relationship between
KIS and R is illustrated graphically in figure 4.1.
There is one further calculation that must be considered
in the quantitative application of the Kubeika-Munk equation. This
involves the reflection at the air-object interface. Whenever a beam of
light strikes a surface of different refractive index, some of the incident
light is reflected from this interface. The reflectances shown for the pig-
ments in appendix A (all glossy samples) are total reflectances designated
as %RT; the surface reflection is included. When the light incident on a
glossy surface is perpendicular (normal to the surface) or near perpen-
dicular, the amount of light reflected at the interface, k, is related to the
refractive index of the material at the interface by the Fresnel equation
(eq. 4.3), which was introduced earlier:
where n
1
is the lower refractive index, often air, and n
2
is the higher
refractive index of the material at the interface.
This is the same equation described at the beginning of this
chapter for transparent materials when the angle of incidence is near to
the perpendicular. The k
1
is the amount of light reflected at an angle
equal and opposite to the incident angle. For air, n
1
is 1.000. If the
refractive index at the interface material is 1.500, which is an average for
organic vehicles, then k
1
(the first-surface reflection) on glossy samples is
0.04, or 4% of the incident light. This is called the specular (or mirror-
like) reflection. If the incident light were totally diffuse, instead of being
perpendicular to the surface, k
1
would be equal to 9.2% (Judd 1942). If
the surface were perfectly diffuse, k
1
would also be 9.2 %. However, we
seldom see such perfectly diffuse surfaces. Due to their multiple surfaces,
pigmented filament fibers may be glossy and exhibit a high degree of
surface reflection when woven, resulting in specular reflection at many
angles that approaches diffuse reflection. Pressed-powder white standards
used with color-measuring instruments may be nearly ideally diffuse.
Their surface characteristics depend on the skill of the preparer and on
the equipment used. Near-white matte paints made with rutile Ti0
2
can
be highl y diffuse reflectors because the amount of white is above its criti-
cal pigment volume concentration. In this situation, the interface surface
contains almost no binder. The high refractive index of the Ti0
2
provides
highl y diffuse reflection. Examples of the effects of surface reflection
characteristics on color and appearance are presented in chapter 6.
The importance of the surface correction for light exiting a
paint or plastic interface as well as entering this material is discussed
80 Chapter 4
later in this section. As a simple approximation, 4% may be subtracted
from the total reflectance measured at each wavelength for the curves
shown in appendix A. The resulting value can be designated with the
subscript NS, meaning nonspecular reflectance or simply diffuse reflec-
tance. It is this lower value of reflectance that is used to calculate the KIS
values that are used in this single-constant approximation. It is also the
value that would be obtained if these glossy samples were measured with
the specular reflection excluded. Kubeika-Munk theory does not take
into account the surface reflection. It is based on the assumption that
there is no interface; hence, this correction must be made if the total
reflectance has been measured.
The first step in the quantitative application of single-
constant Kubeika-Munk calculations is the quantitative determination
of the unit K/S values of the pigments of interest. In order to do this, dry
pigments must first be dispersed, that is, ground into the vehicle used.
Artists' tube colorants are predispersed, so thi s step is not necessary.
Generally, predispersed pigments exist in the vehicle as more finely
divided particles than the pigment particles in dispersions made by hand,
for example, with a muller.
Since the individual pigments are to be calibrated relative to
one particular white, the absorption for the white is simply obtained by
measuring the reflectance of the white film at complete hiding: Kw =
(1 - R)
2
/2R, where R is the decimal fraction of the diffuse (nonspecular )
reflectance. When Kw is known, the unit (K/S)p for each absorbing pig-
ment can be calculated from the reflectance of a film (at complete hiding)
for each pigment when mixed with the white pigment by the following:
(4.10)
where the subscript M refers to the pigment-white mixture, the subscript
W to the white, and the subscript P to each absorbing pigment. C is the
decimal fractional relative concentration, such that Cw is equal to 1.000
minus the total concentration, Cp, for the absorbing pigments .
To even partially define the spectral curve, the reflectance
must be determined for at least 16 wavelengths, for example, every 20
nm from 400 to 700 nm. The unit (K/S)p values at each wavelength are
specific for the particular vehicle system and white pigment used. These
unit values for the pigments of interest constitute the pigment calibration
library for future use. Once created, given good quantitative preparation
of the samples and good measurement repeatability, the experiments
leading to this library need not be repeated. For maximum accuracy, the
measured reflectance for the pi gment-white mixtures at the wavelength
of maximum absorption should be about 40%, the value for the most
precise measurement on a spectrophotometer (Stearns 1969) and for the
greatest sensitivity to change in reflectance with concentration (Johnston-
Feller and Bailie 1982b).
Using a single wavelength only, an example will serve to illus-
trate the pigment calibration just described. In appendix A, the curve for
Colorant Mixture 81
the phthalo blue-white mixture in figure A.10 can be used, along with
the curve for the masstone rutile Ti02 from figure A.5. At the absorption
maximum of the mixture, 610 nm, the nonspecular reflectance (RT - 4%)
nearest 40% is that for 0.25% blue and 99.75% white, that is, Rr=
39.8%. At the same wavelength, the nonspecular reflectance for the
white is 91.0%. From the tables of K/S versus R values in appendix B,
the K!S values for these reflectances are (K/S)w = 0.00444 and (K/S)M =
0.455. Substituting these values into equation 4.10, the unit KIS value
(for 100%) at 610 nm for the blue is
( )
0.455-0.9975(.0044)
K/S = = 180.4
p .0025
From other similar pigment-white mixtures of known con-
centration, a library may be established. With the library of unit K/S val-
ues established, the determination of the concentration of an unknown is
routine. Equation 4.10 rearranged to solve for an unknown pigment con-
centration in a white mixture becomes
Cp = (K/s)M -Cw(K/ s)w
( Kj S )P
(4.lla)
where the subscript M refers to the measured mixture. The concentration
of the white is unknown, but as can be seen in the example of the phthalo
blue above, the actual value of Cw is only insignificantly different from,
and can generally be set equal to, unity. Because the absorption is so low,
the quantity (K/S)w can be subtracted for a first approximation without
introducing appreciable error. (In the case of rutile Ti0
2
, wavelengths
below 430 nm, where this pigment begins to absorb significantly, are
generally avoided in simple approximations of pigment concentration.)
A more accurate estimate of concentration, even using single-
constant Kubeika-Munk theory where Sw is set equal to 1.000, requires
the use of all 16 or more wavelengths, a formidable task without a com-
puter program. However, much can be learned without the use of such a
program, if a knowledge of absorption mixtures is considered. As in the
example above of a single pigment mixed with white, only the wave-
length of maximum absorption need be considered. Thus, some simple
estimations of concentrations can be made using only a few judiciously
selected wavelengths; this is a valuable and rapid procedure.
When the examination of a measured reflectance curve sug-
gests that two or more chromatic colorants in mixture with white, hav-
ing absorption maxima in different wavelength regions, have been used
to make the color, the estimate of the concentration of each pigment
requires the use of reflectance val ues at the maximum absorption region
of each. Thus, if a blue is mixed with a yellow to make a green, the
wavelength of the long-wavelength absorption maximum for the blue and
the short-wavelength absorption maximum for the yellow would be used.
The amount of white is obtained by difference. Since it absorbs only
slightly, its concentration, Cw, can be set equal to unity unless the color
82 Chapter 4
of the mixture is very dark or very high in chroma. (Indeed, the low
absorption of the white is often ignored completely for a first estima-
tion.) If the assumption is made that the white absorption is unity, the
concentrations of two colorants in a mixture can be calculated from the
following two equations:
(4.llb)
(4.1 lc)
where the subscripts 1 and s refer to the long and short wavelengths; M
refers to the mixture; and the numbers refer to the individual pigments.
The K/S values can be determined from the reflectance values by use of
equation 4.7 or from the table in appendix B.
An example of the calculation of two chromatic colorants
mixed with white will illustrate the use of this technique. Consider the
orange represented by the solid line in figure A.30. If the identifications
and quantities of pigments were not known, the examination would be
carried out in the same way as for any unknown: look for absorption
maxima (reflectance minima) beginning in the long-wavelength region
and moving toward shorter wavelengths for other clues. The first absorp-
tion maximum (reflectance minimum) is found at 560 nm. This is the
green wavelength region where red pigments absorb. So the curves of red
pigments in our limited library can be consulted. BON red is quickly
identified as having a single absorption maximum in this region (fig.
A.14). Moving to shorter wavelengths, a maximum can be found at 450
nm in the violet, where yellow colorants absorb. There is little clue to
identify the yellow except the wavelength region where the absorption
decreases; this is just beyond 450 nm. From our limited library this can
only be Permanent Yellow FGL (fig. A.20) . The very short wavelength
absorption is due to rutile Ti0
2
(fig. A.5).
In order to estimate the quantities of the red and the yellow,
the absorptions at the two wavelength maxima identified for both-560
nm for the red and 450 nm for the yellow-can be used. But before any
calculations can be made on the unknown, it is necessary to determine
unit absorption coefficients for each of the identified pigments (the BON
red and the Permanent Yellow FGL) as well as the rutile Ti0
2
white at
these two wavelengths. Referring to figure A.14, the reflectances for
1 % BON red-99% rutile Ti0
2
can be read from the curve at these
two wavelengths; similarly, using figure A.20, reflectances for 2.5%
Permanent Yellow FGL at the same wavelengths can be read. The rutile
Ti0
2
is starting to absorb more at 450 nm, so a correction for the small
amount of white can be made. Table 4.2 summarizes the total reflec-
tances, RT, read from the curves; the reflectances corrected by subtract-
ing 4% for the surface reflectance, designated RNs; the corresponding
KIS values read from the table in appendix B; and the unit absorption
coefficients calculated using equation 4.10. The pigment concentrations
used in the calculations are decimal fractions so the unit K!S values are
for unity (100%).
Table 4.2
Total reflectance (Rr ), read from the spec-
tral curves; reflectances created by sub-
tracting 4 % from the surface reflectance
(RNsl; the corresponding KI S values (from
appendix B); and the unit absorption coef-
ficients (calculated with eq . 4.10).
Colorant Mixture 83
Pigment A.
RT RNS
KIS (K/S)corr
Cone.(%)
(K/5)100
Ti0
2
560 95.0 91.0 0.0044 100 0.0044
Ti0
2
450 93 .0 89.0 100 0.00608
BON 560 46.0 42.0 0.401 0.397 39.7
BON 450 57.5 53 .5 0.202 0 .196 19.6
FGL 560 95.0 91.0 0.0044 0 2.5 0
FGL 450 38.0 34.0 0.641 0.635 2.5 25.4
The calculation for the unit concentration of the FGL is
illustrated by:
(
K! S) = [0.641 - 0.006]= 25.4
100% 0.025
The reflectances for the orange may now be read from the
curve in figure A.30 at the two wavelengths: at 560 nm, RT= 39.0 (RNs =
35.0); and at 450 nm, RT= 33.5 (RNs = 29.5). The absorptions at each
wavelength are the sum of the K/S values contributed by the fraction of
each pigment present. When the color obviously contains a large amount of
white, the exact concentration of the white is not needed to get an estimate
of the amounts of the absorbing pigments present because its absorption is
so low. In the case of the orange example, then, just two equations-one
for each wavelength-are required to estimate the concentrations of the
two absorbing pigments. The equations are of the following type:
(4.lld)
where (K/S)
0
is the absorption of the orange mixture based on the
reflectance corrected for the 4 % surface reflection, and subscript R refers
to the red pigment; subscript Y refers to the yellow pigment; and sub-
script W refers to the white pigment.
Substituting the KIS values of the above reflectances into the
equations for each wavelength gives the following:
at 560 nm: 0.604 = CR(39. 7) + Cy(0.00) + (0.004)
at 450 nm: 0.842 = CR(19.5 ) + Cy(25.5) + (0.006)
The two equations are easily solved because the absorption of
the yellow at 5 60 nm is zero:
0. 604 = CR(39.7) + (0.004)
CR= 0.0151 , or 1.5 %
Now that CR is known, the concentration of the yellow is calculated
from the equation for 450 nm:
Cy= [.842 - 0.006 - .0151 (19.5)]/25.5
Cy = 0.0213, or 2.13 %
Thus, our estimate using two equations is not quite in agreement with
the concentrations, but it is close for a first trial. The reasons that it does
not agree perfectly are several; the variance is primarily due to their
84 Chapter 4
concentrations having been calculated originally using the two-constant
Kubeika-Munk equation (eq. 4.8), and to corrections for internal surface
reflection, k
2
, as well as the correction for external surface reflection, k
1
(eq. 4.12b). The use of these equations is discussed in the next section.
The important point of this illustration is not that the results
are not identical but that they are close enough for a trial that, if the
concentrations and preparations of colorants are reasonably constant,
would yield a preparation that is close in hue and strength and could
easily be corrected to produce the desired color. Additionally, the results
did not require the use of any elaborate computer program. Even if one
has access to such a program, it is important to understand what is
happening in the program and to have the ability to interpret quickly
whether the computer results are reasonable.
Up to this point no mention has been made of the metric
used to specify the concentrations. The unit used to describe concentra-
tions of the pigments in our illustration is relative weight of dry pigment
in wet paint. It could just as well be volume of pigment instead of
weight. Weight or volume, wet or dry, any unit may be used provided it
is used consistently. It is not necessary to know the dry colorant concen-
tration if a colorant is made using only one colorant that primarily
absorbs, or one pigment that primarily scatters. In the case of artists'
tube colorants, the concentration of the dry pigment is not known. If
there are no other absorbing or scattering materials in the mixture of
pigment and vehicle, the concentration of the undiluted wet mixture is
set equal to unity, and fractions are calculated on this basis. The pig-
ments present should be identified on the tube. If more than one pigment
is present, the technique of setting the unknown concentrations to uni t y
is not valid. A single pigment calibration would not suffice; instead, a
series of preparations at different concentrations in white would have to
be prepared. An attempt could then be made to identify the absorption
regions of each individual colorant in the mixture.
This same basic technique can be extended to mixtures of
three or more absorbing pigments; wavelengths representing individual
pigment absorption maxima should be selected judiciously, and the equa-
tions should be solved simultaneously.
An alternative technique for calculating mixtures of two or
more absorbing pigments with white is an iterative method using the fol-
lowing equations, which make the assumption at first that at the wave-
length of maximum absorption, the other component does not contribute
significantly. Thus,
(K I s)
at A. ,: C1 = M (4.lle)
(KI s)
1
where A.
1
is the wavelength of absorption of the major component (desig-
nated by subscript 1). After estimating C
1
from this equation, the val ue
for C
1
is substituted in the following equation to calculate the concentra-
tion, C
2
, of the second component at A.
2
:
Colorant Mixture 85
(4.llf)
Once the concentration of the second component, C
2
, is determined, this
value is substituted in the recalculation of the concentration of the major
component, C
1
, at A.
1
:
(4.llg)
This series of calculations-the iteration-is repeated, using equations
4.1 lf and 4. llg until equation 4.9 is satisfied. (The concentration of the
white may or may not be added into equation 4.9, depending on the pre-
cision desired. It is often of little significance compared with the accuracy
with which the pigment mixtures can be made.)
The techniques described above for calculating opaque col-
orant mixtures using single-constant Kubeika-Munk equations for pig-
ment mixtures and for dyes on fabrics, or for calculating transparent
colorant mixtures using the Beer-Bouguer equation, are the simple tech-
niques used in analytical spectrophotometry. They are as valid today as
they were fifty years ago.
Using a few selected wavelengths for analyzing the spectral
curves of mixtures is a shorthand method of carrying out a complete
spectral curve analysis, sometimes called a differential spectral curve
analysis. In a complete analysis, the curve of the mixture is broken down
into the absorption curves of the individual components that additively
combine to make the absorption total. In the field of color science, the
pioneering computer for carrying out a curve analysis was the analog
Colorant Mixture Computer, COMIC, designed and manufactured by
Davidson and Hemmendinger (Davidson, Hemmendinger, and Landry
1963). The introduction of this simple tool began the use of visible spec-
tral curve measurement and analysis for industrial color control in manu-
facturing processes. The operator of the computer observed 16 dots on
an oscilloscope screen; these dots represented the KIS values for 16
equally spaced wavelengths. By adjusting the amounts of KIS from one to
five components, the curve dots could be zeroed to balance the K!S of the
sample in question. On occasion, the author has used it for some samples
measured by transmission in the infrared region as well. Differential
spectral curve analysis can be useful in any region of the spectrum, and
it is the basis for all analytical interpretation of spectral curves.
Use of the Kubeika-Munk single-constant theory (using K/S
instead of separate K's and S's) is presented by Billmeyer and Saltzman
(1981). These authors follow this illustration wi th a picture of COMIC
(the analog computer) and a flowchart illustrating the iterative procedure
used in computer colorant matching. Kuehni (1975) illustrates a digital
computer colorant-matching program.
86 Chapter 4
Basing the reference library on a single colorant-white mix-
ture was, and is, not generally sufficient for accurate analytical work.
Instead, a range of concentration mixtures was made, that is, a calibration
curve of KIS versus concentration was prepared for each absorbing mate-
rial, and the calibration closest to that estimated to be present was used in
the final calculation. Today, with computers, this may also be done, and a
range of unit concentration values may be stored. However, as in the case
of mixtures, if the scattering may come from more than the white in a
pigment mixture, the more complete two-constant Kubeika-Munk equa-
tion is generally used. The examples presented in this monograph were
calculated using two-constant theory, where separate values for K and S
for all pigments are used in the calculation. In addition, surface correc-
tions for internal reflections are also used.
More complex estimation
In the estimation procedure described earlier in this chapter, two
significant simplifications are made. The first is that the only surface
reflection correction necessary is the subtraction of the first-surface
reflectance from the total measured reflectance. However, before light
that has entered the paint film can exit the film, it encounters another
interface-the paint-to-air interface-where part of the exiting light is
reflected back into the film. The same reflections occur with all dielectric
materials-plastic, glass, paper, etc. The second simplification is that all
the scattering comes from the white pigment or from some other low-
absorbing, scattering component. In many situations, particularly with
dark or high-chroma colors, this additional assumption is not adequate.
Several such cases are pointed out earlier in this chapter (see the section
on "Absorbing and Scattering Materials" ).
The total reflectance measured is described by the following
equation for dielectric materials (nonconductors of electromagnetic
energy, that is, nonmetals):
(1-k1 )(1-k2 )Ri
RT =k1 +------
l-k2Ri
(4.12a)
where RT is the total decimal reflectance; Ri is the internal decimal
reflectance (see eq. 4.12b); k
1
is the decimal amount of the first-surface
reflection coefficient (air-to-paint interface); and k
2
is the decimal amount
of the internal surface reflection (paint-to-air interface).
Ri is used in the Kubeika-Munk equation and is obtained by
rearranging equation 4.12a so that
(4.12b)
The assumption is made in both equations that the light is totally diffuse
internall y, striking the interface in equal amounts at all angles. Equation
4.12a was fi rs t described by Walsh (1926), but it is now generally called
the Saunderson correction because Saunderson first pointed out its
importance when usi ng the Kubeika-Munk equation (Sa underson 1942).
Figure 4.2
lnterfacial reflectances as a function of the
refractive index ratio, n/n
1
, where n
2
is
the denser medium (from Judd and
Wyszecki 1963. Copyright 1963
John Wiley & Sons, Inc. Reprinted by
permission).
Colorant Mixture 87
If the refractive index of the paint at the interface is 1.5, then
k
1
is 0.04 (4%), and k
2
is 0.60 (60%). Saunderson found that using k
2
equal to 0.4 ( 40%) gave the best results on his polystyrene samples.
Experience has shown that it really does not make much difference,
within experimental error, whether 40% or 60% is used, provided that
for any pigmented system, one is consistent and uses the same value for
all pigmented samples, including those used in preparing the pigment cal-
ibration library (Giovanelli 1956; Brockes 1960). For paints, the author
has used the higher, theoretical figure for k
2
-0.6 ( 60% ). More complete
discussions of the best value of k
2
may be found in papers by Mudgett
and Richards (1973) and by Phillips and Billmeyer (1976).
The relationship between refractive index and the surface
reflection is described by Judd (1942; Judd and Wyszecki 1963). Judd's
curve is reproduced here as figure 4.2. A table of exact values of k
1
and
k
2
was computed and published by Billmeyer and Alman (1973).
Another diagram in Judd and Wyszecki's books (1963,
1975), reproduced here as figure 4.3, illustrates why k
2
(the internal
exiting reflection loss) is so high. It shows how light rays are refracted
when entering a medium of different refractive index. When entering a
medium of higher refractive index, the rays are bent toward the perpen-
dicular to the surface; when exiting from a medium of higher refractive
0.50
g
Q)
0.40 u
c
.!'!
u
Q)
;:;:::

<ii
c
0.30
.So
....,
c
Q)
u
<;
CL
0.20
Relative index of refraction, n2/n1
88
Figure 4.3
Illustration shows how light rays emanat-
ing from point P inside a material of
refractive index 1.5 (n
2
) are bent when
entering a less-dense medium, such as air
(n
1
). Because the exiting rays are bent
away from the perpendicular, about 60%
of the light is reflected back into the mate-
rial. The light striking the surface at point
C (42) travels to point 0, and at angles
greater than 42, the light is totally
reflected back into the dense medium.
The 42 angle is the critical angle when
the refractive index of the denser material
is 1.5 (from Judd and Wyszecki 1963.
Copyright 1963 John Wiley & Sons, Inc.
Reprinted by permission) .
Chapter 4
~
~
~
~
I :t
I
/
~ ....
\
I
/
. ~
\ /
\ /
\ /
/
0 0
p
index to a lower one, such as air, the rays are bent away from the per-
pendicular so much that many of them are totally reflected back and
cannot escape. The angle at which light striking the interface from inside
begins to be totally reflected is called the critical angle, and it depends
on the refractive index. For a refractive index of 1.5, this angle is about
42 from the perpendicular.
Figure 4.2 shows that the higher the refractive index of the
vehicle, the larger will be the reflection at the interface for light both
entering and exiting the film. The relationship between %RT and %Ri for
a refractive index of 1.5, based on equation 4.12a, is shown graphically
in figure 4.4. Applying these surface corrections is extremely important
for very dark colors and for colors of very high chroma, where in some
regions of the spectrum the reflectance is very low.
The second simplification made earlier in this chapter is that
all of the scattering comes from the white pigment.
In the case of the yellow-black mixtures, which make an
olive green, it was pointed out that this color results because the yellows
have their maximum scattering in the green region. And even though a
mixture may contain significant amounts of rutile Ti0
2
, the scattering of
the yellows plays an important role (see figs. A.19 and A.21). If we were
to use the full two-constant Kubeika-Munk equation (eq. 4.8), this
behavior would be predicted.
The full two-constant equation relates the KIS of the total
mixture to the sum of the fractional concentrations of each pigment, Cp,
times each corresponding unit absorption coefficient, Kp, di vided by the
sum of the fractional concentration of each pigment, the Cr's, multiplied
by its corresponding unit scattering coefficient, Sp, as follows:
C1K1+C2K2+C3K3+ .. .
C151 +C
2
5
2
+C
3
S
3
+ .. .
where the subscript numerals identify the individual pigments, and the
C's are relative concentrations such that their sum is unity (100%). The
(K/S)M is determined from the calculation of the internal reflectance, Ri,
from the measured total reflectance from equation 4.12b. If diffuse
Figure 4.4
Relationship between total reflectance and
internal reflectance calculated from eq.
4.12 for perpendicular incidence and a
refractive index of 1.5.
Colorant Mixture 89
70

v
u
c
60
.s
u
v
;:;:::
l':'
(tj
50
c

c

40 c
v
::!
v
CL
30
20
10
10 20 30 40 50 60 70 80 90 100
Percent total reflectance (RT)
reflectance is measured instead of total reflectance, the numerator on the
right side of the equation becomes simply RT. The assumption is made
that the paint film is opaque, completely hiding the substrate at all wave-
lengths. Solving this equation for unknown concentrations using 16
wavelengths and 4 appropriate pigments becomes a formidable task
without the use of a computer.
It is not the purpose of these discussions to dwell on the
complex computer color-matching technique using the two-constant
Kubeika-Munk equation, nor to present actual examples. However, a
reader interested in making color matches may wonder how the utiliza-
tion of equation 4.8 is carried out. The first step involves the determina-
tion of the individual pigment characteristics-the K's and S's.
Since there are two unknowns, Kp and Sp, two equations are
required to determine the unit values of Kp and Sp at each wavelength. In
order to calibrate each pigment, that is, to describe its unit Kand S for
each wavelength, two mixture relationships, or two equations for known
mixture concentrations, are necessary. For example, the reflectance
measurements of a known mixture with a standard white and a second
known mixture with a standard black give the two equations needed for
simultaneous solution. The procedure requires a standard white and a
standard black. A further quantity that can be of value in pigment
calibration is the reflectance of the masstone of a single pigment that
90 Chapter 4
contains no black and no white. Thus, there is a choice among the fol-
lowing three equations for obtaining the unit Kr and Sr values.
For white mixtures: ~ J M
CpKp+CwKw
CpSp +CwKw
For a masstone (eq. 4.7): (1 - R; )
2
/2R; = Kp/Sp
For black mixtures: ~ J M
CpKp +C
8
K
8
CpSp +C
8
K
8
(4.13)
(4.14)
Note that all reflectances used in the Kubeika-Munk equation are
internal reflectances obtained by use of equation 4.12a rearranged as
equation 4.12b.
From the two pairs of equations-equation 4.13 with equa-
tion 4.7; or equation 4.14 with equation 4.7-two sets of K's and S's can
be computed. The K's and S's of the pigment in the absorption region of
the spectrum are selected using results obtained from equation 4.13 for
the white mixture and equation 4. 7 for the masstone. The K's and S's of
the pigment in its scattering region are selected from the wavelengths of
its maximum reflectance using results obtained from equation 4.14 for
the black mixture and equation 4. 7 for the mass tone. The values of the
K's and S's in the crossover region between the two sets of data are
selected on the steepest part of the reflectance curve between minimum
and maximum by smoothing the values; generally there is a wavelength
where the two sets of values are very close. In the case of transparent
pigments, when mixtures with black are near black, all of the values used
are from the white and the masstone, equations 4.13 and 4. 7. This pro-
cedure for determining the K's and S's used for computing colorant for-
mulations was first described by Davidson and Hemmendinger (1966).
If a masstone cannot be prepared at complete hiding, its reflectance may
be calculated from films measured over white and over black using the
equations for incomplete hiding described below under the heading
"Opacity, Translucency, and Hiding Power."
It must be remembered that all of the unit K's and S's are
relative to the particular white used in the calibration procedure. If a dif-
ferent white is desired or necessary, its K's and S's must be calculated rel-
ative to those of the original white. For example, if rutile Ti0
2
had been
used for the calculation of the unit K's and S's of the pigments, but a cal-
ibration for ZnO white is desired, the K's and S's for the ZnO must be
determined relative to the K's and S's for the rutile Ti0
2
This is most
readily done from a mixture of the ZnO with the standard black and a
ZnO masstone with equations 4.14 and 4.7.
A method for obtaining the K's and S's for the standard white
and the standard black pigments remains to be described. There are two
methods for determining the K and S for the white. The first of these is a
relative method and the second, an absolute method. The relative method
for determining the K and S of the white involves only the preparation of
a masstone white that is completely opaque. The scattering coefficients,
the S's, are set equal to 1.000 at all wavelengths, and the absorption
Colorant Mixture 91
coefficients, the K's, are calculated with equation 4.7 (the equation for
the masstone). The K's and S's for the standard black are then calculated
relative to these values for white by use of equations 4.13 and 4.7, that
is, from a white-black mixture and the masstone black. This relative
method works well as long as the samples are at complete hiding.
When hiding is incomplete, a different version of the
Kubeika-Munk equation must be used (Kubeika 1948; Ross 1967):
R = 1-Rg(a-bctgh bSX)
' a - Rg + b ctgh bSX
(4.15)
where R; is the decimal internal reflectance of the film over the substrate
of internal reflectance, Rg.
b = a
2
-1
( )
l/2
R
0
= reflectance over a black of 0% internal reflectance
Rco = internal reflectance at complete hiding
(4.16)
(4.17)
x =thickness (actually ex, since altering the concentration, c,
has the same effect as altering the thickness, X, by the same
multiple)
ctgh = the hyperbolic cotangent
Tables of hyperbolic cotangents are available, for example, in Judd
and Wyszecki (1975), which also gives equations for calculating hyper-
bolic functions.
An equation useful for determining the scattering
coefficient S is
1 [
1
[ 1- a Roll
SX = b ctgh - bi;;
(4.18)
where ctgh-
1
is the inverse hyperbolic cotangent. From these equations it
can be seen that S can be determined from the values of a calculated from
reflectances, R;, measured over a light substrate of reflectance, Rg and R
0
,
or calculated from R
0
and Roo by use of the appropriate equation from
those in equation 4.16. In either case the thickness, X, and concentration,
C, must be known. The value for K is calculated from the relationship a=
(S + K)IS. The absorption and scattering coefficients thus calculated are
dimensioned, being in terms of the thickness, X, and the absolute concen-
tration, C. (The absolute concentration is the volume or weight percent-
age in the dried paint film or in the solid plastic piece, for example.)
The K's and S's of the black and of the individual colorants
are subsequently determined as before; now the absolute values of the
white are substituted for the relative values used when all S's were set
equal to unity in the complete hiding calculation. These pigment calibra-
tions made relative to the absolute values of the standard white may now
92 Chapter 4
be used to calculate hiding power or to predict Re" from samples at
incomplete hiding. If one wishes to calculate the colorant composition
for a glaze to be applied over another color, these absolute values are
needed (see " Glazes" below). If one wishes to determine the absolute
amount of pigment necessary to achieve a desired degree of opacity
(Osmer 1978), these values are also needed (see "Opacity, Translucency,
and Hiding Power" below). Not all of the colorant-formulation com-
puter programs avai lable include this type of calculation, which is so
very useful in many situations involving incomplete hiding layers.
After determining the unit K's and S's at 20 nm intervals or,
preferably, at 10 or 5 nm intervals, the colorant formulation can be cal-
culated to match the color desired. This is described in terms of the
reflectances at the same wavelengths as those used for the pigment cali-
brations. The amounts of individual pigments are determined by means
of simultaneous equations. How the match is achieved may vary with
the particular computer program, but, in general, an iterative calcula-
tion is carried out by use of a variety of pigments designated by the
operator. The initial goodness of match may be estimated by a least-
squares calculation of the reflectances; a visually weighted least-squares
calculation of the reflectances; or the calculation of tristimulus values
and color differences.
After calculating each possible match, most computer color-
matching programs select those matches that are least metameric by cal-
culating the tristimulus values for other selected illuminants and then
calculating the color differences for each from the desired color. Thus, if
a match has been made for illuminant C, such that t:.E is zero, the color
differences for several other illuminants, such as illuminant A (incandes-
cent light source) and a fluorescent lamp (often abbreviated F) are then
calculated. These differences are an approximate measure of the degree
of metamerism (see the section on "Metamerism" in chap. 2).
Use of the computer color-matching program is not limited
to applications involving color. Computer programs for two-constant
color matching may not be as versatile as those for single-constant, but
they can be useful in other regions of the spectrum. Osmer (1982) illus-
trated their usefulness in computing pigment formulations in the near-
infrared region as well as in the visible region for paints to be used for
camouflage applications. (Incidentally, with the matches for the very
dark green, he illustrates that computer matches can be calculated that
contain no white pigment but depend on chromatic colorants for pig-
ment scattering.)
Although digital computer color-matching programs have
many applications, they are not a substitute for intelligence nor a
panacea for sloppy control and handling of the components. If the opera-
tor offers the computer twelve pigments to use four at a time, he will
often get a very large number of matches. However, a knowledge of the
spectral curve shapes of pigments enables the intelligent operator to nar-
row down considerably to the most likely or the most suitable pigment
choices at the beginning, saving much time and paper.
Colorant Mixture 93
Applications of Kubeika-Munk Formulas
Study of colorant changes
One important application of the Kubeika-Munk equations, which is of
particular interest in the museum world, is in the study of colorant fad-
ing or darkening. Knowledge of the rate of these changes of the colorants
used in materials enables the museum personnel to develop the best con-
ditions for the preservation of the objects in their care while still allow-
ing them to be displayed.
Basically, establishing the rate of fading, or darkening, is an
analytical problem involving measurement of the change in concentration
of the colorants. Thus, use of the Kubeika-Munk equation for the deter-
mination of concentration changes can provide the analytical data neces-
sary for studying the types and rates of change that occur during aging
or following exposure to deleterious conditions.
In the introduction to this monograph, it was pointed out
that our eyes are not analytical but are integrative, in terms of the stimu-
lus that forms the basis of our perception of what we see. The CIE
method of describing color, and any derived color-difference equations,
represents an attempt to provide a numerical description of what we see.
Thus, the use of such calculated data is inappropriate for analytically
describing the changes in the concentrations of the coloring agents.
Color plate 1 illustrates a series of seven different mixtures
of alizarin with rutile Ti0
2
Compare the series on the left with that on
the right. Can one tell by visual evaluation the fraction of the original
alizarin concentration in the left column of color chips that remains in
the corresponding color chips on the right? Not at all. The fact is that
the color chips in the right column illustrated in color plate 1 all have
lost 30% of the original alizarin concentration present in the correspond-
ing chips shown in the left column.
In applying Kubeika-Munk calculations to determine the rate
of colorant change, the simplest situation will be described first: evaluat-
ing the concentration change of a specific single pigment mixed with
white in a particular vehicle, exposed in a specified manner. (In the case
of the evaluation of dye fading, the analogous simple situation is evaluat-
ing the concentration of a single dye applied to a colorless substrate in a
specified manner.)
Many pigments, with the exception of some of the highly inert
inorganic pigments, vary in stability when exposed to light, heat, high
humidity, or extreme values of pH, depending on the surrounding medium
(the vehicle). Therefore, it is important that the stability be described in
relation to the particular pigment-medium (vehicle) system. The application
to be described involves a known single pigment, or single pigment-white
mixture, exposed to whatever deleterious factors are considered important.
Accelerated-aging exposures may also be used, if one remembers that these
conditions do not necessarily represent the real world exactly.
In the limited application to be discussed, the test samples
prepared for evaluation are simple mixtures with white, in the case of
94 Chapter 4
pigments, or single dyeings in the case of dyes, prepared for the specific
purpose of the study. For this simple application, one does not even have
to know the exact concentration. One simply makes the assumption that
the white pigment does not absorb significantly in the region of the
absorption maximum. The reflectance at this wavelength is measured,
and the nonspecular reflectance converted to K/S. The quantity (K/S)
0
for the unexposed sample is the reference to which the subsequent (K/S)e
values for the exposed samples are compared as follows:
100[ (KI S). / (KI St] =% colorant remaining
(4.19)
The subscript e refers to the exposed sample and the subscript o to the
unexposed reference. The K!S values are based on the diffuse reflectance.
If the original concentration is known, it may be multiplied by the
fraction remaining to obtain the actual concentration remaining. This
example is an illustration of the simple estimation of colorant concen-
tration described above in the section "Quantitative Application of the
Kubeika-Munk Equation." The rate of change can then be interpreted
by application of the principles used in the study of the kinetics of
chemical reactions.
In this initial discussion the assumption is made that the pig-
ment fades to a form that does not absorb in the region of the original
absorption maximum. Many pigments do this. Some organic reds, such
as synthetic alizarin lake, fade through a yellow form in the progress of
becoming completely colorless. However, the yellows formed do not
absorb in the region of the absorption maximum of the red form, so they
do not interfere with this estimate of concentration remaining (Johnston-
Feller 1986). Figure 4.5 illustrates the fading of alizarin following expo-
sure in the Atlas xenon arc Fade-Ometer.
A few pigments darken instead of fading to a colorless
form. Vermilion (the red form of mercuric sulfide) is one of these. Other
inorganic pigments that darken include some that contain lead, such as
chrome yellow and molybdate orange; Emerald Green (copper acetate-
copper arsenate) also may darken. Feller (1967) published spectrophoto-
metric curves showing the darkening of vermilion in oil without added
white pigment. Some of those curves are shown in figure 4.6. Studying
the curves, one can see that two changes are occurring: ( 1) at the long-
wa velength region of maximum reflectance, the reflectance is decreasing
due to the darkening; and (2) at the short-wavelength (absorption)
region, the reflectance is increasing as the concentration of the red ver-
milion decreases. Feller discusses the rate of darkening as measured by
the change in KIS in the long-wavelength region.
Studies of pigment change can be profitably analyzed by
use of a color-matching program, utilizing the multiple wavelengths
described above (in the "Quantitative Application of the Kubeika-Munk
Equation" section) and the full two-constant equation, equation 4.8.
When computer color-matching programs are applied to the study of
color changes following exposure of a single colorant-white mixture, it is
necessary to use calibrations for four pigments (including white) in order
Figure 4.5
Reflectance curves of alizarin in mixture
with rutile Ti0
2
, initially and following
exposure in the Atlas xenon arc Fade-
Ometer. The exposure is expressed in
terms of kilojoules per square meter as
measured at 420 nm (after Johnston-
Feller 1986).
Colorant Mixture 95
- V--l...,lf-- B ------- G
y 0 -----R----
100
90
80
70
}g
60
g
<!)
u
c
.s
u
50
<!)
q::

...,
c
<!)

<!)
40 0..
30
20
10
400 450 500
alizarin lake
17.6% in rutile Ti0
2
550
Wavelength (nm)
2368 kJ m-2
unexposed
600 650 700
to have the necessary three degrees of freedom in color space. Thus, in
the case of a single pigment in mixture with white, two other pigment
calibrations must be included in the calculation. The three must be
selected so that one controls the redness-greenness, one controls yellow-
ness-blueness, and one controls lightness-darkness. In the case of the cal-
culations for the alizarin-Ti0
2
illustrated in figure 4.5, calibrations for a
yellow pigment and for a black pigment were used in addition to those
for the alizarin and the rutile Ti0
2
In this way, fading, yellowing, and
darkening could be determined (Johnston-Feller et al. 1984). By use of
this technique, it was shown that synthetic alizarin pigment in mixture
with white faded through a yellow intermediate before fading to a color-
less form, and the amount of the yellow intermediate was quantified.
A further reminder concerning such studies must be empha-
sized. The samples prepared for study must be at complete hiding and at
such thickness and pigment volume concentration that hiding remains
96
Figure 4.6
Reflectance of vermilion (cinnabar) before
and after exposure in the Fade-Ometer. It
can be seen that the color is approaching
the very dark gray characteristic of
metacinnabar.
Chapter 4
G y 0 -----R----
<ii
.....,
g
Q)
u
c
fl
u
Q)
<;::::

.....,
c
Q)
u
Q;
CL
100
vermilion in oil
90 --- original data
80
70
60
50
40
30
20
10
400
- - - 171 hours exposed in Fade-Ometer
433 hours exposed in Fade-Ometer
450 500
/
/
/
550
Wavelength (nm)
/
,...-
---
/
/
/
. . .. .
600
..............
. . . .
650 700
complete over the time of the study. Pigment deterioration increases as
light penetrates more deeply into the film and fading continues. For this
reason, measurements on the samples should be made over white and
over black. When hiding becomes incomplete so that the black backing
can be seen, the simplified Kubeika-Munk equation (eq. 4. 7) is no longer
valid. The dependence of the depth of fading on the pigment volume con-
centration of the Ti0
2
and on the relati ve alizarin concentration in the
alizarin-Ti0
2
samples made at constant thickness is illustrated in color
plate 2. The height of the dowels corresponds to the depth of fading. The
dowels were painted with the actual material used in the samples (see
figs. 9 and 10 in Johnston-Feller 1986).
Changes in color perceived as fading or darkening can result
from reasons other than deterioration of the pigment. For example, if the
surface reflection characteristic of a paint changes over time, the appear-
ance may seem to indicate a change in colorant concentration. The gloss
Figure 4.7
Transmission electron micrograph of a
cross section of a fluorocarbon plastic pig-
mented with Ti0
2
(large black particles)
and with carbon black (clouds of fine black
dots). The arrows point to voids that sur-
round the pigment clusters.
Colorant Mixture 97
of high-gloss samples may decrease, resulting in an increase in the sur-
face scattering and thereby giving the appearance of fading. In the case
of low-gloss (matte) materials, the gloss may increase because of burnish-
ing, fingerprints, washing, addition of varnish, and so on, resulting in a
perceived darkening of the color. Whether the exposed sample is lighter
or darker, the changes in the spectral reflectance curve due to changes in
the surface reflection are different from the changes due to pigment dete-
rioration: reflectance changes due to surface changes tend to occur uni-
formly at all wavelengths, resulting in a nearly constant increase or
decrease in reflectance across the spectrum (see remarks on chalking of
paint, Johnston and Feller 1967).
Spectrally uniform changes in reflectance also result from
increases or decreases in light scattering that occur inside the coating,
such as at the pigment-vehicle interface. If the suspending vehicle
(medium) matrix shrinks over time, for example, the result may be the
formation of microscopic voids surrounding the pigment particles. Such
voids scatter light very effectively, creating the effect of added white pig-
ment and lightening the perceived color (see "Microvoids and Vesiculated
Beads" in chap. 6). Thus, though the appearance may be that of fading
(loss of colorant), the actual cause can be quite different.
Figure 4. 7 shows a transmission electron micrograph of an
exposed fluorocarbon plastic pigmented with Ti0
2
and carbon black.
Carbon black and Ti0
2
incorporated in a stable plastic matrix (vehicle
or medium) normally do not exhibit fading. Yet this sample appeared to
have faded. Careful study by electron microscopy of the cross section of
the exposed material revealed that voids exist around pigment clumps in
many regions (disregard the streaks in this picture-they are caused by
the microtome slicing). The voids contain no vehicle and no pigment;
they are virtually empty cavities, possibly filled with air if it has diffused
into them. The voids have a much lower refractive index than the sur-
rounding material. Hence they scatter light, increasing the perceived
reflectance. Such a change inside the film cannot be detected by spectral
analysis as being different from fading (loss of colorant) .
98 Chapter 4
Note also in this electron photomicrograph that the concen-
tration of the Ti0
2
particles (large black particles) and of the tiny, black,
dot-size carbon black particles is not uniform throughout the plastic
matrix. As is apparent, the pigment volume concentration (PVC) is low,
which is typical in thick plastics; the pigments tend to clump together and
are not uniformly distributed throughout the material. This occurs in
many paint films made at low PVC. For example, Buttignol (1968) shows
electron micrographs of paint samples made at PVCs of 0.3 to 1.3 (gener-
ally low for paints). The uneven distribution of pigment can be observed
in his figure 5 for transparent synthetic yellow oxide and in his figure 9
for transparent red oxide, even though the dry pigments were fairly uni-
form in particle size. The causes of this nonuniformity could include lack
of wetting of the pigment particles by the vehicle, so that van der Waals
forces hold the pigment particles together; the presence of moisture sur-
rounding the pigment particles; or the interaction of the pigments with the
vehicle (some pigments tend to inhibit the curing of the vehicle).
The lack of wetting can be serious in other cases. For example,
if paint used to retouch areas of a painting does not adhere well to the
original, the added paint layer may separate from the original over a
period of time. The air present between the layers causes light scattering
and an increase in reflectance across the spectrum.
If the pigment particles are not thoroughly wetted in the
initial pigment dispersion, darkening may also result during exposure.
Perhaps the fluid decomposition components in the vehicle are able to
wet the pigment more than the "fresh" vehicle in the original dispersion.
This would result in a significant change in transparency as well, because
of the lower degree of scattering. Feller (1968) contrasts the absorption
KIS curves of two greens, both of which darkened on exposure: Emerald
Green (copper acetate-copper arsenate), which darkened because of pig-
ment change in the oil vehicle; and Green Earth in oil, which presumably
darkened because of an increase in wetting of pigment-vehicle interfaces
or interstices. Figure 4.8 (from Feller's fig. 6) illustrates the reflectance
change of Emerald Green following exposure in a Fade-Ometer. From the
reflectance curve shape it is not readily apparent what is happening. But
when the absorption K/S curves are plotted on a logarithmic scale, an
interpretation of the change can be made. Figure 4.9 (from Feller's fig. 7)
illustrates the changes in absorption KIS of the Emerald Green and the
Green Earth before and after exposure.
There are two important points to be emphasized in these
curves. The first is the value of studying log absorption K/S curves
graphically-K/S on semilog paper versus wavelength-that is, learning
to visualize spectral curves in absorption terms rather than simply in
reflectance terms. Log absorption curves retain a constant shape regard-
less of the concentration of the absorber (Derby 1952). Thus, if a pig-
ment appears to decrease in concentration, the change in the spectral log
absorption curve will be the same at all wavelengths, decreasing uni-
formly; conversely, if the pigment appears to increase in concentration,
the scattering may have decreased, or the absorption may have increased
uniformly at all wavelengths (see discussion in the next section).
Figure 4.8
Reflectance of Emerald Green (copper
acetate-copper arsenate) before exposure
in the Fade-Ometer and after exposure
under glass, with and without an ultra-
violet filter (Plexiglas UF-1). Note that the
darkening is less for the exposure under
the UF-1 filter than for the exposure under
glass (after Feller 1968).
Colorant Mixture 99
y 0 -----R----
90
80
70
] 60
g
"'
u
c
<U
1l 50
<;::
1:'
....,
c
"' !:::'
~ 40
30
20
10
400
Emerald Green oil paint
exposed Ill hours in Fade-Ometer
--- unexposed
- - - covered by glass+ UF-1
- - - - covered by glass
-- .....
/ '
/,. .. -- .. ,,
/ .. ' '
I , , '
/
, ~
, ' ~
/..
, .. "'
, ..
/;
~ ..
......... _ ..
-..
450 500 550 600
Wavelength (nm)
650 700
The second point is that when the difference between exposed
and unexposed log KIS curves is not uniform, there must have been a
change in the colorant system, not just a change in colorant concentration.
The curves in figure 4.9 illustrate these points. The log
absorption curves for the exposed and unexposed Emerald Green differ
by unequal amounts at different wavelengths, evidence that a change in
colorant has occurred. Since the maximum change has occurred at the
absorption minimum (reflectance maximum), darkening has taken place
(which can be judged visually as well), as exhibited also by some degree
of increase in absorption at all wavelengths. In contrast, the two curves
for the exposed and unexposed Green Earth are more nearly uniformly
separated, indicating that the primary change is simply an increase in
absorption or, more likely, a decrease in scattering owing to a change in
the pigment dispersion or pigment-vehicle wetting. One can postulate
that this apparent increase in absorption may have occurred because of
100
Figure 4.9
Comparison of two types of darkening for
two different green pigments in oil. The
increased absorption of the Emerald Green
when exposed (decreased reflectance
under the ultraviolet filter illustrated in
fig. 4.8) is greatest at the absorption maxi-
mum at 530 nm and decreases at other
wavelengths, indicating a change in the
chemical structure to a darker form. The
Green Earth, in contrast, darkened uni-
formly at all wavelengths, suggesting an
increase in the wetting of this pigment.
Bleaching of the vehicle accounts for the
lack of parallelism between the KIS values
for the exposed and unexposed Green
Earth oil paint.
Chapter 4

c
0
:;:;
u
c
:J
'+-
c
0
:;:;
5.0
Q._ 1.0

.D
<{
0.5
400
'
'
'
y 0 -----R----
darkening of green
pigments in oil
' /
\ /
\ /
\ I
450
\ I
\ I
\ I
\ I
\ I
\ I
\ I
\ I
\ I
', /
....... _,,

......... _.
. . .
--- unexposed Emerald Green
- - - exposed Emerald Green
unexposed Green Earth
exposed Green Earth
1.778 X unexposed
500 550
Wavelength (nm)
600 650 700
increased refractive index of the linoxyn (oxidized linseed-oil vehicle); or
because of increased wetting of the pigment, possibly by low-molecular-
weight degradation products in the vehicle, with a subsequent decrease in
scattering due to air initially entrapped in and around the pigment part-
icles. Feller suggests that "this is a type of alteration [decreased scatter-
ing] that may lead to pentimenti [a perceived increase in transparency]."
The smaller change observed at the short-wavelength region, indicating
less absorption, strongly suggests that the oil vehicle has also bleached, a
not-uncommon change when linseed oil is exposed to ultraviolet light.
Instruments can be helpful in diagnosing the nature of per-
ceived changes. Surface change can be measured with a spectrophoto-
meter if the instrument is designed for the inclusion and exclusion of the
specular reflectance. Generally, such an instrument uses an integrating
sphere for collecting the reflected light. The incident light beam is off
the sample perpendicular by about 8-10. At the angle equal and oppo-
Colorant Mixture 101
site to the incident angle, that is, -8 to -10, there is a removable port
in the sphere wall. With a black port or a light trap in place, the specu-
lar reflectance is absorbed and only the diffuse light reflected by the
color is measured. If a white port is inserted instead (one that matches
the sphere wall), all of the light-that is, the total reflectance-is mea-
sured. The difference between these two (total reflectance minus diffuse)
is a measure of the specular reflectance. When measuring exposed col-
ors, both types of measurement should be made whenever possible. This
procedure will demonstrate whether a change in the surface reflectance
(gloss) has occurred.
A procedure for quantifying the changes occurring following
exposure was outlined by Johnston-Feller and Osmer (1977). Their tech-
nique was based on measurements made on a sphere-type spectropho-
tometer with the specular reflectance both included and excluded. A
further step suggested was a gentle washing of the surface to remove the
chalking and any dirt accumulation. With the aid of a computer pro-
gram, the quantification of the different types of changes, calculated in
visual terms, could be made. Analytical data concerning the concentra-
tions of the pigments could also be calculated.
Changes in the surface reflection can also be measured using
a glossmeter, but quantification in visual terms (fiE) cannot be calculated
by this means.
If no computer program is available, the analysis described
for hand calculation earlier in this chapter can be adapted for measure-
ments made with the specular reflectance included and excluded, and by
use of the same techniques described by Johnston-Feller and Osmer. If
there is no area of the object where the color was protected, such as
under the rabbet of a framed painting, there is no way to know what
may have changed or by how much. It is only when there is a reference
point at some time during the exposure that any of the changes discussed
above-fading, darkening, changes in the dispersion, or changes in
gloss-can be quantified.
The logical conclusion is that all objects of great value,
particularly those subjected to long exposure to light and to changes
in environment, should be measured periodically. However, not only
could this be a Herculean task, but many materials, when exposed in
a museum environment, change very slowly, so that the small changes
would scarcely be evident in a single generation, say twenty-five years.
Moreover, in this day of rapidly changing technology, commercial instru-
ment obsolescence takes place in only five to ten years, making the spec-
trophotometric measurements of reflectance difficult to reproduce exactly
over very long periods of time. An elaborate program for instrument cali-
bration and a supply of very stable reference standards would be neces-
sary before a program of detecting small changes in color over long
periods of time could be embarked upon.
The problem of making reflectance measurements repro-
ducible with a view to detecting small changes will be discussed again
in the instrumentation overview provided in chapter 9. In the case of
objects made with fugitive colorants, such as Japanese prints, changes are
102 Chapter 4
sufficiently rapid that the use of readily available methods of reflectance
spectrophotometry can be very helpful (Feller, Curran, and Bailie 1984 ).
In recent years the National Gallery, London, has undertaken
an extensive program to develop the instrumentation and technology to
monitor color change in paintings using modern image processing and
computational capabilities (Saunders 1988; Saunders and Cupitt 1993).
Vehicle (medium) or substrate change
Another important factor in studying works of art-paintings in any
medium and on any substrate, or textiles-is to determine whether
the observed changes in color are due to vehicle changes, to substrate
changes (fabric or paper, for example), or primarily to a change in
the pigment. In general, following exposure, painting media (vehicle)
changes, white paper discoloration, and textile substrate deterioration
are characterized by yellowing and, occasionally, by slight darkening.
With some materials, such as old linseed-oil films, bleaching may occur.
Since changes in absorption involving yellowing are most apparent in
the blue region of the spectrum, they are most significant on the violets,
blues, and purples, which exhibit their major reflection in this short-
wavelength region. On the other hand, the color of yellows, oranges,
and even reds are less visually affected by this increased absorption in
the short-wavelength blue and violet region, where they already absorb
significantly. Thus, vehicle or other substrate yellowing can lead to a con-
siderable distortion in the overall color balance originally present in a
museum object. These basic principles were pointed out by Laurie (1 926)
in his groundbreaking book The Painter's Methods and Materials.
An example from Johnston (1967b) may make this effect
more clear. Two unsaturated polyester plastic panels containing white
and a single chromatic pigment were exposed outdoors in Florida. After
six months of exposure, both panels had changed slightly with a t:i.E
of 1.5 CIELAB units, several times more than a minimum perceptible
amount. One panel was pigmented with 5 % of a green-shade cobalt blue
with 95% rutile Ti02, the other with 20% cadmium orange with 80%
rutile Ti02. The curves are reproduced here as figure 4.10 for the cobalt
blue panel and figure 4.11 for the cadmium orange panel. Examination
of the curves for the cobalt blue indicates that the concentration of the
blue pigment, as indicated by the absorption maximum at 640 nm, had
not changed at all. All of the change occurred in the short-wavelength
blue region because of increased absorption due to vehicle yellowing. The
changes in the curves for the cadmium orange (fig. 4.11) are also limited
to this short-wavelength region. Differential spectral curve analysis of
both panels showed that, while all of the change in the blue panel was
caused by vehicle yellowing, the differences in the cadmium orange panel
were attributable to the combination of fading of the cadmium orange
and yellowing of the vehicle. The latter partially offset the loss of about
16% of the orange pigment originally present. The amount of yellowing
due to the vehicle was about the same in both panels. The analysis of the
cadmium orange panel following exposure was difficult because both the
pigment and the dirty yellowing of the plastic absorb in the same wave-
Figure 4.10
Curves of a colored plastic made with
cobalt blue pigment and rutile Ti0
2
before
and after exposure outdoors in Florida for
six months. No change in the concentra-
tion of cobalt blue has occurred, as can
be seen at the absorption maximum at
610-640 nm. All of the color change is
caused by yellowing of the plastic vehicle;
this is shown by the decrease in the
reflectance in the violet-blue region of
the spectrum.
Colorant Mixture
- v--11-c---- B -->--- G y 0
Q)
u
c
90
80
70
-8 50
Q)
==

....,
c
Q)
40
CL
30
20
10
400
I
I
I
I
450
5% cobalt blue
500 550 600
Wavelength (nm)
103
----- R----
650 700
length region. The two exhibit differently shaped spectral curves, how-
ever. Use of differential spectral curve analysis techniques with COMIC
provided an approximation of the amounts of each factor present, the
concentration of orange remaining, and the absorption curve of the yel-
low component that was formed (the yellow could be quantified arbi-
trarily as a real yellow pigment with a similar curve).
Figure 4 .12 shows the reflectance curve of a yellowed varnish
on white gesso, measured with the specular reflectance excluded. The cor-
responding values of KIS are shown on the right ordinate of the graph.
This illustration of a transparent but absorbing layer applied over a reflect-
ing substrate can be considered that of a glaze, although it is not often
thought of as such. Glazes are discussed more fully in the next section.
The important point in illustrating increased yellowing upon
exposure, which is so characteristic of deterioration, is that, in general,
changes in the substrate, such as paper or a textile, look very similar in
104
Figure 4.11
Curves of a colored plastic made with cad-
mium orange pigment in Ti0
2
before and
after exposure outdoors for six months in
Florida-the same exposure period as that
of the cobalt blue illustrated in figure 4.10.
The amount of yellowing of the vehicle
was the same as for the cobalt blue
sample, but because the cadmium orange
faded, the visual color change was little
noticeable. Differential curve analysis
showed that 16% of the cadmium orange
had faded, but the resulting loss of
absorption of the cadmium orange was
counterbalanced by vehicle yellowing,
which increased the absorption in the
short-wavelength region.
Chapter 4
y 0 -----R----
90
20% cadmium orange
80
70
60
50
40
30
unexposed
20
10
400 450 500 550 600 650 700
Wavelength (nm)
change of curve shape to that for the yellowing of a traditional natural-
resin varnish such as damar. Figure 4.13, for example, illustrates the
effect of adding lignin to paper. The curve shape changes are not identi-
cal to those of the yellowing curve in figure 4 .12 for damar varnish, but
the general shape is that of yellowness and slight darkening, which is
characteristic of deterioration of colorless or white organic compounds
such as vehicles (media) or substrates (paper or textiles). Learning to
read these characteristic changes in terms of the spectral curves can
be extremely helpful in diagnosing the type of change that may have
occurred following exposure.
If a computer color-matching program is available, it can be
used to great advantage when the changes following exposure are stud-
ied. By this technique it was found that in white mixtures at complete
hiding, alizarin faded to a yellow form before becoming colorless; a ye!-
Figure 4.12
Reflectance curve of a yellowed varnish
(damar). The absorption in terms of KIS is
indicated on the right.
Colorant Mixture
y 0
90
yellowed, darkened varnish
80
70
60
v
u
c
.s
u
v
<;:::
50
.....
c
v
l:'
v
Cl..
40
30
20
10
400 450 500 550 600
Wavelength (nm)
----R---
105
KIS
0.010
0.025
0.050
0.075
0.10
0.15
0.20
0.30
0.40
0.50
0.60
0.80
1.0
2.0
3.0
4.0
10.0
650 700
low component had to be added to match the exposed samples
(Johnston-Feller and Bailie 1982a). Computation to match the exposed
Emerald Green (figs. 4.8 and 4.9) indicated that black had to be added to
create the match. However, computation to match the exposed Green
Earth (illustrated in fig. 4.9) indicated that Green Earth had to be added
and that yellowed vehicle had to be subtracted. Such computer color-
matching techniques are simply another method of differential spectral
curve analysis.
An interesting and useful diagnostic tool for museum conser-
vators concerned with color-balance distortion due to deterioration yel-
lowing was described by Lafontaine (1986). He proposed that one look
at the object illuminated by an auxiliary light source filtered to provide
extra blue light and less yellow light, in order to restore the reflected light
distribution as it would have been if the yellowing of the varnish or
106
Figure 4.1 3
The effect of increased amounts of lignin
coated onto a white filter paper. Although
not identical in curve shape to the yel-
lowed damar of figure 4.12, the nonselec-
tive spectral absorption decrease with
increasing wavelength is roughly similar-
both are dirty yellowing curves (after
Johnston and Feller 1967).
Chapter 4
u
<ii
u
100
90
80
70
8 60

50
c
.s
u
Q)
q::
1:'
c 40
Q)
u
Q:;
"-
30
20
10
400
hardwood lignin coated
onto filter paper
--- filter paper
- - - 0.5%
1.0%
2.0%
y 0
450 500 550 600
Wavelength (nm)
-----R----
650 700
vehicle had not occurred. When examining paintings using this technique,
the amount of yellowness compensated for can be adjusted by use of a
dimmer on the fi ltered auxiliary light source. This method of viewing
deserves to be used widely by conservators and curators when examining
objects where yellowing from deterioration is suspected- on paintings,
drawings, decorative painting, printing on paper, and textile samples.
Glazes
The applications of the Kubeika-Munk equations discussed so far have
been based on the assumption that the samples are opaque, that is, the
substrate is completely obscured and does not contribute to the observed
color (the condition of complete hiding). As mentioned, however, the
problems of yellowed varnish, vehicle, or substrate are not limited to
samples at complete hiding. Multiple layers of pigments applied over one
another at incomplete hiding may be encountered in the study of many
Colorant Mixture 107
museum objects and materials. The term glazes, as used here, refers to
the situation in which one or more chromatic layers do not completely
obscure the substrate-that is, the circumstances do not correspond to
complete hiding. According to this definition, most watercolors, some
printing inks, and some textiles fit into this category, in addition to paint
layers at incomplete hiding-the traditional glazes.
Kubeika (1948) described the hyperbolic mathematical solu-
tions (eqs. 4.15-4.17) to the equations for the original Kubeika-Munk
model (Kubeika and Munk 1931) described earlier in this chapter. To
review, equations 4.15-4.17 are
R
' a-Rg +b ctgh bSCX
where Ri =reflectance (internal) over substrate Rg
Rg = reflectance (internal) of the substrate
a = 1 + KIS
b = a
2
- 1
( )
11 2
ex= absolute concentration times thickness
ctgh = hyperbolic cotangent
For a mixture in the incomplete hiding layer of thickness X,
[
C1K1 + C2 K2 +C3K3 + ]
a = 1 + X ---------
C1S1 +C
2
S
2
+ C
3
S
3
+
(4.15)
(4.16)
(4.17)
(4.20)
Additional equations are described in the appendix of an article on glazes
by Johnston-Feller and Bailie (1982a) and in the next section of this
chapter. These more complex equations point out that the perceived
color and the spectral reflectance curve are different for ( 1) the case of
two colorants applied at incomplete hiding over each other, compared
with (2) the case of the same two colorants physically mixed together
and applied at complete hiding.
Johnston and Feller (1967) published the reflectance curves
for four pigments-two yellows and two reds-that differed in scattering;
they were made up as glazes over rutile Ti0
2
and in mixture with the
same Ti0
2
. Also included in their examples were the curves for the mas-
stones and for the glazes applied over black made at the same thickness
and concentration as the glazes over white. The curves are reproduced
here as figures 4 .14-4 .1 7.
Figure 4.14 shows chrome yellow light as an example of a
highl y scattering, largely opaque pigment. Figure 4.15 for Flavanthrone
Yellow is an example of a low-scattering, transparent yellow. The differ-
ence in scattering is apparent from the curves for the masstones and for
the glazes applied over black. The masstone chrome yellow reflects highly
and is a bright yellow; the masstone Flavanthrone Yellow reflects much
less and is a dirty, low-chroma yellow. The glazes applied over black
exhibit the scattering characteristics strikingly. Note that the reflectance
108
Figure 4.14
Reflectance curves of an opaque
chrome yellow pigment mixed with
rutile Ti0
2
applied at complete hiding;
a masstone applied at complete hiding; a
masstone applied at incomplete hiding
over rutile Ti0
2
; and the same thickness
applied at incomplete hiding over black.
Note that the reflectance over black is
plotted at 5 times the measured
reflectance.
Chapter 4
G y 0 ----- R----
100
Regal Chrome Yellow LT
(I mperial X-2777)
over white
90
/.-----------
mixed with white /.
--
80
70
60
50
40
30
20
10
400
.
I
.
I
.
I
.
I
.
I
.
I
.
I
.
I
' I
! /;
/I ,'
i' I
,/ ,'
/I I
'/ I
J ,'
/ ,'
//
'LI
I
I
/
I
'
;
/
/
;
'
'
'
i '/
_/
-------7
.-
450 500 550
Wavelength (nm)
--
'
......
'- ...._over black
'
'
.....
5x ...._ ....
--
600 650 700
curves for the latter are multiplied by 5 in the plot. Chrome yellow scat-
ters much more light than the Flavanthrone Yellow. Note that the resul-
tant color in each case is a green, because of the greater scattering in the
green wavelengths relative to the scattering in the yellow, orange, and red
longer-wavelength region.
Figure 4.16 illustrates the reflectances of comparable paint
samples prepared from cadmium red medium, a moderatel y scattering
pigment. Figure 4.17 illustrates the reflectances of samples prepared from
Quinacridone Red, a relatively low scattering, transparent, red pigment.
The masstone colors depend on the relative scattering: cadmium red is
a moderatel y bright red, while the quinacridone pigment, also a red,
is relativel y dark. (Many coats of paint were required to prepare the
Quinacridone Red masstone at complete hiding. )
Comparison of those mixtures with white to the glazes
over white, for all four pigments, illustrates a basic principle: glazes over
Figure 4.15
Reflectance curves of a transparent organic
yellow pigment mixed with rutile Ti0
2
applied at complete hiding; a masstone
applied at complete hiding; a masstone
applied at incomplete hiding over rutile
Ti0
2
; and the same thickness applied at
incomplete hiding over black. Note that
the reflectance over black is plotted at 5
times the measured values. The slight dip
at Jong wavelengths in the Ti02 mixture is
due to a minor contamination by phthalo-
cyanine blue (see chap. 7 and fig. 7.7).
Colorant Mixture 109
100
90
80
70
60
50
40
30
20
10
400
y 0
Flavanthrone Yellow
(Harmon Y-5713)
-----R----
over white
./------
/ mixed with white
. ------- ........ __
I
. ,
I I
I
I ,'
i ,'
I
I I
I
I I
I
I ,'
I
I I
I
I I
I
! ,'
/ ,'
I
,,
I
J
;
,
; .
" . ,
, "
,'I
; .
.,--.. ., I .... __ ,,,,, ,,
....
/,-""

....--=--/
. --
' ..... over black
---
------ - -
5x
450 500 550 600 650 700
Wavelength (nm)
white result in higher chroma than can be achieved by mixture with
white in all cases, whether the pigment is a high or a low scatterer.
Artists have used glazes for their unique optical effects for
many years. When the substrate is not white but chromatic, many differ-
ent color effects can be achieved. In appendix A, the greenish color
achieved when black is mixed with yellow (figs. A.19 and A.21) is illus-
trated. However, when a scattering yellow, such as the chrome yellow
light of figure 4.14, is applied as an incomplete hiding film, or glaze, over
black (or some other dark substrate), the optical effect is unique-eerie
and tenuous in substance. What fun an artist can have!
An important conclusion from a comparison of the mixtures
over white (the glazes) with the mixtures with white (discussed in an ear-
lier section) is that the characteristic absorption patterns are present in
both. The two transparent organic pigments, Flavanthrone Yellow (fig.
4.15) and Quinacridone Red (fig. 4.17) clearly illustrate this constancy in
110
Figure 4.16
Reflectance curves of an opaque cadmium
red pigment mixed with rutile Ti0
2
applied
at complete hiding; a masstone applied at
complete hiding; a masstone applied at
incomplete hiding over ruti le Ti0
2
; and the
same thickness applied at incomplet e hid-
ing over black. Note that the reflectance
over black is plotted at 5 times the mea-
sured reflectance.
Chapter 4
100
90
80
70
60
50
40
30
20
y 0
Cadmium Red Medium
(Harshaw No. 100)
--- - -R----
over white
,,,. .
-
. / ;
I ,';
. ,
I ,'
I ,'
/ mixed
/
/
with white
I
! ,'
I ,'
, I
I I
' I
/ '
' I
/ ,'
.,
,,
;:
' ,.
,f
, .
----- ____ .,,I
,"' --------- .
I ~ / ,'
I _,
over black
10
I ------------ I
, /" /
400
/ . -- -- -- """'
- - - - - - - - - ..... --
450 500 550
Wavelength (nm)
600 650 700
the characteristic absorption bands. Note the two bands of the flavan-
throne (a distinctive pattern when encountered in a yellow colorant) and
the three bands for the Quinacridone Red, also unique. Whether the pig-
ment is used as a glaze or as a mixture with white, the characteristic
absorptions remain.
A yellowed varnish is, in effect, a yellow glaze over the entire
painting. Removing this glaze results in a considerable change in the
observed color balance. Frequently controversy may arise as to whether
something else-an intentional colored glaze, for example- might have
been removed as well as the yellowed varnish. Analyzing the differences
in the spectral curve shapes before and after varnish removal can some-
times be helpful. Johnston and Feller (1963) published an illustration of
such a spectral curve analysis for an exposed red-orange, glossy paint
panel. The reflectance curves of the panel before and after varnish
removal are reproduced as figure 4.18. Inspection of the reflectance
Figure 4.17
Reflectance curves of a transparent organic
red pigment mixed with rutile Ti0
2
applied
at complete hiding; a masstone applied at
complete hiding; a masstone applied at
incomplete hiding over rutile Ti0
2
; and the
same thickness applied at incomplete hid-
ing over black. Note that the reflectance
over black is plotted at 5 times the mea-
sured reflectance.
Colorant Mixture 111
- v---11-c----- B ------- G
y 0 -----R----
90
Quinacridone Red
(Du Pont R T - 796)
over white
/. - . :-: ,:-::.-::-.:
. ,, ..
/ ' mixed

1
1
with
80
/ / white
70
I
I ,'
I ,'
i II
I
w /1
' I
,'r" ! II
Ii \ I 11
50
1
1
, \ I
1/ \ /1
I ' 'I
,I ~ /1
40 '! ~ . i,1'
\.. .
I . ', f
30 ....... , I
\' .. _, .... ,
. ~ -'
20
10
'/ .. I
. \...
masstone
5x
over black
---------
400 450 500 550
Wavelength (nm)
600 650 700
curves reveals that after varnish removal the curve is lighter and bluer.
Beginning the inspection at the long wavelengths, the increased reflec-
tance in the red region indicates that some absorber has been removed,
possibly dirt. The increased reflectance in the green region, around 540
nm, indicates that some absorption has been removed here, where a
red might absorb, in addition to absorption loss observed at the long-
wavelength region. The large increase in reflectance in the violet region,
around 440 nm, indicates that a yellow has been removed. This analysis
raises questions: ( 1) Was the increased reflectance in the green region due
to a loss of red from the original painting? (2) Was this apparent loss of
red due to a red-tinted glaze that was removed? (3) Was a yellow glaze
removed? Observation of the swabs used to remove the varnish indicated
that a red had been removed, but with just this evidence, its identifica-
tion could be made only by solution pigment identification techniques
(Johnston and Feller 1963).
112
Figure 4.18
Reflectance curves of a red area of paint
before and after removal of a yellowed
varnish.
Chapter 4
If the reflectance curves in figure 4.18 are converted to the
absorption function, K/S, the absorption curve of the material removed
can then be obtained by subtraction (differential spectral curve analysis).
Figure 4.19 illustrates the results (Johnston and Feller's original fig. 5).
Note that the KIS absorption values are plotted on a log scale so the
shape of the curves is independent of the concentration. The difference
curve (curve 1) shows two absorption bands at about 560 nm and 520
nm, present in both the before and after curves. Absorption at these
wavelengths is characteristic of carmine. Subsequent removal of another
layer, also red, resulted in a color with a different red curve, one with a
single absorption band, which could be attributed to alizarin. One sus-
pects, therefore, that a carmine glaze had been applied over the alizarin
(a red over a red). For reference, the absorption curves of the yellowed
varnish (fig. 4.12) and of alizarin and carmine are also included.
Differential curve analysis is thus a useful tool to aid in pigment
100
90
80
70

::J
60


Q)
u
c
_;g
50
u
Q)
""

...,
c
Q)
u
40 (;::;
c..
30
20
10
400 450
G
y 0
before removal
500 550
Wavelength (nm)
600
----- R----
650 700
Figure 4.19
Spectral absorption curve 1 shows the dif-
ference between the reflectance curves in
figure 4.18. Alizarin (curve 3), carmine
(curve 4), and a typical yellowed varnish
(curve 2) are also illustrated.
Colorant Mixture 113
- V ----c--- B ------- G
y 0 -----R----
Vi

c
0
:;::;
u
c
..2
c
0
:;::;
Q..

..0
<(
1.0
0.1
1 - - - difference between curves measured before
and after removal of varnish
2 typical yellowed varnish
3 - - alizarin
4 carmine
\
.. '
.... '
. '
... '...... 1
.. ..... _____ ,
... ......._,
\ \
\, 3 ....... --.,. \
\ /. \ \
. . \
\/ .
/\ \ \
. . \ \
. \
\,
\\
. \
\,
\\
. \
\,
,,
.,,,
0.01
400 450 500 550
Wavelength (nm)
600
.
. '
' ......
. , '
.,
..
.
650 700
identification and quantification, and in other analytical applications of
interest to the conservation of museum objects.
Earlier in this chapter it was pointed out that type II chro-
matic pigments, those that absorb much more than they scatter, when
mixed with white, exhibit a critical concentration at which maximum
chroma occurs. This behavior is illustrated on the chromaticity diagram
in figure 2.12. All of the samples illustrated in the figure were paints
applied at complete hiding.
In the case of glazes, there is a critical concentration where
maximum chroma is reached for an incomplete hiding film applied over
a reflecting substrate for all chromatic colorants, types II and III. In this
case the concentration is actually the product of the concentration times
the thickness, or CX. Thus, for glazes, such a critical concentration exists
for type III colorants, those that both scatter and absorb significantly, as
well as for type II pigments.
114 Chapter 4
The effect of this phenomenon is illustrated by a series of five
concentrations of alizarin glazes applied at incomplete hiding. The glazes
were used to paint the sleeves of a blouse in a portrait done on a Mylar
sheet that was then exposed over a white background for increasing
periods of time in a xenon arc Fade-Ometer (Johnston-Feller 1986). At
intervals during the exposure, the portrait was removed, the glazes were
measured over white, and photographs were taken under carefully con-
trolled conditions. The bodice of the blouse was made with a stable blue
pigment that did not fade. The photographs are reproduced in color
plate 3. The black-and-white graph of the Munsell Chroma versus the
Munsell Value illustrates that the more-concentrated glazes (those of
lowest Munsell Value) increased in Munsell Chroma as the concentration
of alizarin decreased from fading: samples 1, 2, and 3. In the course of
fading, sample 3 (the middle concentration studied) reached maximum
Chroma, exhibiting the critical concentration for alizarin glazes over t his
white substrate. Visual inspecti on of the photographs illustrates the
effect of the phenomenon-a severe loss in the modeling of the drapery
of the sleeves.
In summary, one cannot tell with certainty from the spectral
reflectance whether a pigment has been mixed with another pigment or
applied as an incomplete hiding film, or glaze, over the other. Both are
examples of subtractive colorant mixtures, revealing the characteristic
absorption bands of each pigment. If the film is applied at incomplete
hiding over a white, the contrast in the absorption bands is enhanced.
Undoubtedly there are also many instances when incomplete hiding lay-
ers are applied over opaque chromatic substrates. Even if the mode of
mi xing is not known, whether by mixture with white or other chromatic
colorants or by means of incomplete hiding films or glazes, the character-
istic absorption bands are still visible to aid in the identification of the
pigments used.
Opacity, translucency, and hiding power
Up to the discussion of glazes in the previous section, all of the applica-
tions of the Kubeika-Munk equation described were based on opaque
materials, that is, subst ances that completely hide the background or
substrate. Many materials do not fit this restricted category, however.
While glazes represent a special type of translucent (incomplete hiding)
colorant layer, a wide range of other materials present problems where
their degree of opacit y or translucency has an effect on their appearance.
In thi s section, other applications of the Kubeika-Munk incomplete hid-
ing equation (eq. 4.15) will be discussed.
De"finitions
The terms used to tell whether or not a substrate is hidden must be
defined and used carefully. Definitions in common language dictionaries
are often not precise enough. One must distinguish between absolute (or
quantitati ve) and relative (or qualitative) terms, and among processes,
Colorant Mixture 115
measured quantities, and properties. The following descriptions are based
on ASTM Standard E 284, "Standard Terminology of Appearance."
The property of opacity, which is the ability of a sample to
prevent all transmission of light, is described by the adjective opaque,
meaning transmitting no light. Both of these are absolute terms: either a
sample is opaque or it is not; there is no intermediate state. It is incorrect
to describe a sample as "more (or less) opaque" than another.
The noun transmission and the verb transmit describe the
process whereby light passes through a sample. These terms are relative:
a sample may transmit more or less light than another. The quantity of
light transmitted is measured by the transmittance, which is the ratio
of the transmitted light to the incident light. Thus transmittance is also
a relative term. The transmittance of samples that do not scatter light is
a measure of the amount of light they absorb.
If a sample transmits light without scattering or diffusion, it
is said to be transparent, which is an absolute term. However, the related
term transparency has a special meaning in appearance measurement; it
is described in the discussion of plastics near the end of this section.
Samples that do scatter or diffuse light, unless they are
opaque, are said to be translucent. This is a relative term, since some
samples scatter more light than others. Translucency implies that,
because of a sample's scattering or diffusion of light, objects located
beyond it cannot be seen clearly through it. Of course, translucent
samples may absorb as well as scatter light.
Like transmission, reflection and absorption refer to
processes. Like transmittance, reflectance and absorbance refer to mea-
sured quantities. All of these are relative terms. But recall that, while
reflectance is defined like transmittance as a ratio (reflected light to inci-
dent light), what is always measured is the reflectance factor, that is, the
ratio of the light reflected from the sample to the light reflected from the
white standard. However, for simplicity, we use the term reflectance in
place of reflectance factor in this monograph.
The third term used in the titl e of this section is hiding
power, which is the relative degree of hiding of the substrate. This rela-
tive term as normally interpreted refers to the reflectance measured over
a black substrate (designated R
0
) divided by the reflectance over a white
substrate (designated Rw)-that is, R
0
/ Rw. This relationship is called the
contrast ratio. Discussed at greater length later in this section, this ratio
describes materials that are not at complete hiding.
Even when materials appear to be opaque-that is, the mate-
rials appear to be the same color when observed over a black and over a
white substrate-there may be a perceptible appearance difference
between them in the characteristic we call translucency. In this situation,
the materials exhibit more or less depth of the light penetration into the
material, a subtle, but important, visual appearance characteristic. Some
materials are perceived to have a greater depth of light penetration than
others. Unless we can section such materials horizontally, in order to
decrease the thickness so that we can measure differences over white and
116 Chapter 4
black substrates at various levels within the material, it is difficult to
characterize this aspect of translucency numerically.
Opacity calculations
In the simple Kubeika-Munk equation (eq. 4.7) for samples at complete
hiding
the ratio KIS is dimensionless because the absorption coefficient, K, and
the scattering coefficient, S, have the same units. However, when hiding
is incomplete, the K and S do not appear as a ratio, and their dimensions
must be defined. The following equations and procedures are used t o
determine dimensioned scattering and absorption coefficients.
(4.21 )
where S is the scattering coefficient, e is the absolute concentration of
the pigment, X is the measured thickness of the film, R
0
is the internal
reflectance measured over black, and R,,, is the internal reflectance of a
completely hiding, opaque layer of the same material.
Rx= a - b (4.22)
where a and b are factors based on reflectance measurements and calcu-
lated using the equations seen here.
The calculations also require equation 4.16
a = _!_ [R + R o _ - _ R _ _ R ~ l
2 1- R
0
Rg
and equation 4.17
b = (a2_
1
)112
To determine S, a is first determined based on measurements
made over a black substrate, R
0
, over a white or near-white substrate, R,
and on the near-white substrate itself, Rg, using equation 4.16. The thick-
ness and concentration of the samples over the near-white or white sub-
strate and the black substrate should be identical. From the value for a
thus obtained, b is calculated from equation 4.17. Once a and b are
known, R,,, is calculated from equation 4.22. The value for sex can then
be calculated from equation 4.21. The unit value for S is then determined
by dividing sex by the known absolute values for e (the concentration
of the colorant ) and X (the thickness of the sample). The scattering
coefficient, S, is thus dimensioned in terms of the units used for e and X.
To determine K, the following equation can be used:
K = S(a - 1) (4.23 )
K then has the same dimensions as S.
Colorant Mixture 117
If the thickness is expressed in mils (0.001 in.), the coefficients
K and S are for a 1 mil thickness. The concentration may be expressed in
terms of volume, such as cubic centimeters of scattering material per cubic
centimeter of solid material (for example, the pigment volume concentra-
tion, PVC); or in terms of weight: grams of pigment per gram of pigment
plus cured vehicle (pigment weight concentration, PWC).
From these absolute absorption and scattering coefficients so
determined, the reflectance can be calculated at any desired concentration
and thickness. The incomplete hiding reflectance over any substrate,
chromatic or achromatic, can be calculated from equation 4.15:
R
1
a-Rg +b ctgh bSCX
where ctgh is the hyperbolic cotangent. Tables of the values of ctgh are
given by Judd and Wyszecki (1975).
It must be remembered that most of the reflectance values
used in the Kubeika-Munk equations are internal reflectances. Proper
corrections for the surface reflection should be made using the
Saunderson correction.
When pigment mixtures are used, it is not necessary to deter-
mine dimensioned K's and S's for all of the pigments in the library. (In
fact, it is difficult to do so in the case of very transparent pigments applied
over black.) One simply uses the absolute K's and S's for the white pig-
ment as described above, rather than setting all the S's equal to 1.0. The
procedure for determining the K's and S's of the other pigments in the
library is identical to that described for the complete-hiding samples in the
previous section. The absorption and scattering coefficients of the colored
pigments are then in the same units as those of the white pigment.
After determining the K's and S's for all of the pigments in
the library, the reflectance of mixtures at incomplete hiding over any sub-
strate can be calculated by use of equation 4.20 followed by equations
for b (eq. 4.17), for R (eq. 4.15), or for Roo (eq. 4.22).
Many of the variables encountered when working with col-
ored materials may be studied by use of a suitable version of these equa-
tions. For example, when making calibrations on transparent pigments
as described in the previous section, preparing a film that is at complete
hiding (R,,) is often difficult; many layers have to be applied to build up
the necessary thickness. By using the above technique (measurements
over white and over black), Roo can be calculated for use in the equations.
It has been found that a contrast ratio (R
0
/Rw) of at least 0.75 at the
reflectance maximum should be used for calculating Roo reliably using
equations 4.16, 4.17, and 4.22 (Osmer 1978).
One of the difficulties encountered when applying these equa-
tions to highly reflecting white pigments is that a may be less than 1.0.
It is therefore impossible to calculate b using equation 4.17 because one
cannot take the square root of a negative number. Equations described
by Ross (1967) circumvent this problem.
118 Chapter 4
The color or reflectance observed at any given thickness is
that of a cross-sectional area through the thickness of the film; it is
dependent on the volume of pigment particles that the incident light
encountered as it passed through the defined depth of material.
Expressing concentration in terms of volume units is, therefore, most
appropriate.
As was pointed out in the beginning of this chapter, when
describing changes that occur during aging, pigment particles may exist
as agglomerates, not as simple particles, making their volume difficult to
ascertain when they are incorporated into a polymer matrix. In practice,
pigment mixtures are more reliably measured on a weight basis, which is
then converted to a volume basis using the pigment density (weight per
unit volume). Also, as mentioned, in the case of incomplete hiding, the
concentrations must be expressed in absolute terms based on the pigment
volume per unit volume of the total cured material. This is not the same
as the relative concentration unit used in equation 4.8 describing the
complete hiding (opaque) materials, where the total pigment concentra-
tion was considered to be unity.
Scattering, simple and multiple
In chapter 3, in the description of colorant characteristics, it was pointed
out that for pigments, the two optical characteristics of importance
are the absorption and the scattering of light. It was stated there that
absorption is a relatively simple interaction with light, but that scattering
is more complex. The amount and the angular distribution of scattered
light depend in complicated ways on the particle size, refractive index,
and concentration of the pigment. The brief discussion on light scatter-
ing presented here is limited to pigments with the right size to scatter
light strongly-a particle diameter of about half the wavelength of the
light (too small for the individual pigment particles to be visible in a
light microscope)-and a refractive index quite different from that of
the binder or medium.
When the concentration of pigment particles is quite low,
there is a good chance that the light scattered by a particle will find its
way out of the sample without encountering another scattering particle.
This type of scattering is called single scattering and is described by the
Mie (1908) theory. Experiments confirm the theory and show that in
single scattering most of the scattered light travels in the forward direc-
tion-that is, away from the light source. (An example is given in the
next chapter, in the section describing metallic flakes.) In contrast, when
the concentration of pigment particles is high, the light scattered by one
particle strikes another particle and is rescattered. This occurs many
times before the light reaches the front or back surface of the sample;
this is called multiple scattering. When multiple scattering takes place,
the result is that the light is distributed isotropically, that is, equal
amounts travel in all possible directions, in the backward as well as the
forward hemisphere around the particle. One of the key assumptions of
the Kubeika-Munk theory is that the scattering is isotropic.
Figure 4.20
Percent transmittance of films pigmented
with rutile Ti0
2
plotted as a function of
the pigment volume concentration, \/, mul-
tiplied by the thickness, X. In zone 1 the
scattering is in the forward direction,
which decreases the transmittance, as
though the pigment in the film were an
absorber. In zone 2 a transition is taking
place, and the forward scattering is
increasingly replaced by isotropic scatter-
ing. In zone 3 the transmittance does not
change, because all of the scattering is
isotropic: the film is opaque. Zone 1 is
often referred to as the Mie range; zone 3
as the Kubeika-Munk range (after Craker
and Robinson 1967).
Colorant Mixture 119
In beautiful pieces of experimental work, Craker and
Robinson ( 1967) measured the effect of PVC and thickness of rutile Ti0
2
in paint films on the relative amounts of forward and isotropic scatter-
ing. They first measured the regular (undeviated) transmittance-that is,
the unscattered transmitted energy-using a conventional commercial
spectrophotometer. Figure 4.20 presents their data graphically as mea-
sured on a Unicam SP500 spectrophotometer. Note that the ordinate (the
transmittance) is a log scale. The abscissa is the product of the PVC-
symbol V-times the thickness, X, in mils. It can be seen that in zone 1,
at low concentrations of scattering (nonabsorbing) white particles, the
log of the transmittance decreases linearly with increases in concentra-
tion. This behavior, which appears to resemble the Beer-Bouguer equa-
tion (Beer's law) (eq. 4.5c), is described by -log t = VXk, where tis the
transmittance, Vis the PVC, X is the thickness, and k is a constant char-
acteristic of the material. Thus, VX is the volume of the particles. In
zone 2, a transition from forward scattering to isotropic scattering is
occurring. In this area, slight changes in thickness can effect a significant
change in the ratio of the two types of scattering, particularly for VX in
the range of 1.3 to 1.8. In concentration zone 3, there is no unscattered
(regularly transmitted) light coming directly through the film; the scatter-
ing is isotropic, as assumed in the Kubeika-Munk equation.
Craker and Robinson also measured the angular distribution
of energy scattered around a paint film using an experimental goniospec-
trometer. Figure 4.21 is redrawn from their figure 2; it shows typical
angular results plotted on a polar diagram for a thin film, area A, and
for a thicker film, area B. In the illustration, the incident light is desig-
nated 10, the regularly transmitted light is T, the forward scattered
energy is EF, and the backward scattered energy is EB. In area A (the thin
film), the sharp spike opposite the incident light is the regularly transmit-
ted light. The diffusely scattered light is greater in the forward direction
than in the backward direction. In area B, measurements were made on
a thicker film, and there is no regularly transmitted energy, that is, no
spike. Most of the scattering is isotropic.
"'
u
c
fl
E
Vl
c

.....
c
"'
"'
CL
100.0
10.0
1.0
Zone 1
1
Zone 2 Zone 3
2 3 4 5 6
vx
120
Figure 4.21
Angular distributions of light reflected by
rutile Ti0
2
pigmented films. With thin
films, area A appears to be near the transi-
tion area between zones 1 and 2 of figure
4.20. In the thick film (area B), where no
direct transmittance is apparent, the
reflectances are closer to zone 3, the
isotropic scattering zone, in figure 4.20
(after Craker and Robinson 1967).
Chapter 4
lo lo
j
l
Es
thin thicker
EF
T
area A area B
These angular distributions for the thin and thick films can
be associated with the transmittance curve of figure 4.20. The thin-film
reflectances, described as area A in figure 4.21, appear to be near the
onset of the transition area between zones 1 and 2; in the thick film, area
B, where no regular transmittance is apparent, the behavior appears to
be closer to that in zone 3, the isotropic zone.
The phenomenon of the relative amounts of forward (single
particle) scattering to isotropic (multiple particle) scattering is important
when the problem of translucency is considered. It also explains some
examples of seemingly anomalous behavior encountered in the everyday
world. For example, a printer once showed the author a pair of samples
where thin films of white (probably Ti0
2
) were printed on aluminum
foil. The first sample had one coat of the white ink, and the second
sample had two coats. The second sample was darker than the first one.
He was asked what would happen if he put a third coat on the second
sample, and he stated that it would be lighter than when the sample had
two coats. The probable explanation is that the CX of the double coat of
white ink was in the zone 1 region where the scattering is in the forward
direction, but that the addition of the third coat moved the material into
zone 2, where isotropic scattering began to occur. The low diffuse reflec-
tance of aluminum foil (on the order of 20%) contributed absorption of
the forward-scattered light from the first coats.
It is important to remember that the phenomenon of forward
or isotropic scattering with change in pigment concentration is not lim-
ited to white pigments but may occur with any pigment that scatters
significantly (type III pigments as described in chap. 3).
For a theoretical discussion of single and multiple scatter-
ing, the reader is referred particularly to chapter 3 in the book by
Kortum (1969 ).
Figure 4.22
Schematic diagram of the integrating
sphere of spectrophotometers and col-
orimeters with near-normal incident light,
1
0
, striking the sample and reference at an
angle about go off the normal (perpen-
dicular). If the inserts in the sphere wall
are black, or a light trap, the specular
reflectance, at _go, is absorbed, and only
the diffusely reflected light is measured by
a photodetector perpendicular to the plane
illustrated. If the inserts are white, match-
ing the sphere wall, all of the light
reflected by the sample is included in the
measurement.
Colorant Mixture 121
The degree of translucency or the relative opacity of materials
is not an easy characteristic to measure. Obviously, because translucency
is dependent on the scattering characteristics, an instrument with an inte-
grating sphere is necessary. If a material to be measured is not at com-
plete hiding, a number of measurements can be helpful in characterizing
the degree of translucency. Figure 4.22 presents a diagram of the inte-
grating sphere of a typical commercial double-beam spectrophotometer
used for color measurement. There are six openings in the sphere wall
(excluding the receptor opening). The first two openings are for the
entrance of the incident light, one for the sample light beam and one for
the reference standard beam. Nearly perpendicular (8-10 off normal)
to the incident beams, on the opposite side of the sphere, are two more
ports for the measurement of the reflectance of a sample relative to a
standard (usually the white reference), which is in the second port. The
third set of openings is located on the wall opposite these reflection ports
at the angle equal and opposite to the incident light angle, providing for
the inclusion or exclusion of the specular reflection from the sample and
reference standard.
For rigid translucent materials, such as pieces of plastic,
ceramic, and so on, the total transmittance through the sample in the
forward direction may be measured by mounting the sample at the
entrance port. For the corresponding reference beam, either no sample
may be inserted, in which case the reference is air, or a piece of unpig-
mented material may be used. The latter arrangement makes an optical
compensation for the first-surface (interfacial ) reflection. However, the
surface correction at the opposite side, where the light is exiting from
the materials, is not correct, because more surface reflection occurs on
the scattering sample than on the reference clear material. To simplify the
measurement, air is generally used as the reference.
If the reflection ports and the specular inclusion ports are
made white to match the sphere wall, all of the light transmitted in the
ref.
spl.
122 Chapter 4
forward direction is measured. If the sample reflection port is black,
however, the regular transmission-that is, the unscattered light-is
absorbed by the black, and only the light scattered diffusel y in the for-
ward direction is measured. The difference between these two measure-
ments is a measure of the regular transmittance. Note that when the
sample is mounted in the entrance port in this type of measurement, all
of the light scattered in the backward direction is excluded from the
measurement and is reflected back into the instrument.
If one wi shes to determine the amount of light scattered in
the reverse direction-that is, reflected rather than transmitted-the
sample must be mounted in the reflection port on the opposite side of the
sphere, along with a reference standard in the other port, as is normally
done for reflectance measurements. When samples of interest are not
opaque, they must be mounted on an opaque substrate for measurement.
To avoid problems of surface reflection at a rear-air interface, the back-
ing should be opticall y sealed to the sample. (This is achieved by use of a
viscous sealing material of refractive index similar to that of the sample;
this eliminates air between the two materials when the "sandwich" is
made.) If the backing is white, all of the light transmitted or scattered in
the forward direction, as well as the reflected light, is measured. If the
backing is black, only the light reflected backward into the sphere is
measured. The specular ports in the sphere are used for exclusion or
inclusion of the first surface specular reflection: black for exclusion and
white for inclusion.
Designations for these measurements are usually R
0
for the
measurement over black and Rw for the measurement over white. If the
white backing is not perfectly white, that is, R < 100%, and if the black
backing is not a perfect black, that is, R > 0%, corrections can be made
for the actual values so that Rw and R
0
are properly corrected (see Ross
1967). When the substrates are in optical contact with the films, Rw and
Rg are the total reflectances. The total light transmitted and reflected in
the forward direction, that is, TT, can be calculated from these two mea-
surements by
(4.24a)
If the reflectance of the black backing is less than 1 %, but the white back-
ing (substrate) does not reflect 100%, the following equation can be used:
(4.24b)
where Rg is the reflectance of the near-white substrate. The TT deter-
mined by this reflectance technique should be the same as that obtained
by measuring the total transmittance in the forward direction when the
sample is mounted at the entrance port, as explained earlier.
The techniques just described for measuring the reflectance
and transmittance of translucent materials apply onl y to materials that
are not completely opaque. But, as mentioned near the beginning of this
section, opaque materials may vary in appearance in a subtle way as a
Colorant Mixture 123
result of the perception of depth of light penetration into the materials-
that is, their degree of translucency. For example, a problem presented to
the author was to help in designing plastic piano keys to replace the nat-
ural ivory traditionally used. Not only did the gloss, color, and grain of
the ivory have to be simulated, but the degree of translucency had to be
approximated as well. In fact, wider differences in color were found to
be acceptable than differences in translucency. Natural ivory keys were
first measured for color and translucency, using the techniques over black
and white substrates as just described. Subsequently, synthetic plastic
approximations were measured in the same way for comparison. Plastic
simulations were produced that adequately reproduced the real thing.
Results of the measurement work on this project are described in U.S.
Patent 4,346,639 issued to Vagias (1982).
Unless an opaque translucent material is expendable so that
it can be sectioned horizontally, or the thickness can be decreased by
abrasion or dissolution, measurements over black and white cannot be
exploited, and the character of the translucency cannot be calculated
using the Kubeika-Munk equations described.
The depth of penetration of incident light is also a major
appearance attribute of glazes (described in the previous section). It is
not possible to match the appearance, even though the instrumentally
measured color can be matched, unless the substrate and glaze colorants
are applied in a manner similar to that used in the original. One must
not forget that color is only one aspect of appearance.
Among industrial products, white light diffusers, particularly
those used with fluorescent fixtures, represent a practical application of
the principles of translucency described above. In this case, the scattering
of the white particles has to be in the forward direction for greatest
efficiency. Thus the pigment concentration and particle size must be care-
fully controlled.
Should the reader assume that the techniques for measuring
translucency of relatively thick materials are easy to carry out with
accuracy, a word of caution must be mentioned. The measurements are
subject to many variables due to instrument design, even though an inte-
grating sphere instrument is used. In addition, sample size and thickness,
surface reflection, and the substrate backings also affect the measure-
ments. One of the troublesome effects of sample size and thickness is
loss of light scattered out of the edges of translucent sheets. Atkins and
Billmeyer (1966) tested a number of commercially available instruments
and concluded that none of them could "measure the transmittance and
reflectance of relatively thick translucent specimens with high absolute
accuracy." Nevertheless, if measurements are made on one particular
instrument type in the same manner each time, valuable comparative
results can be obtained.
In the case of thin, translucent films in good optical contact
with the substrates, the reproducibility can be very good. The paint
industry, for example, relies heavily OR the measurement of substrate
hiding in terms of the contrast ratio, which is a measurement of reflec-
tance of a film applied over black divided by the reflectance of a film of
124 Chapter 4
the same thickness and pigment concentration applied over white. Hiding
power is discussed at greater length below.
The plastics industry extensively uses the determination
of the opacity of sheets and of thin films. It applies a special term to
describe the scattering in translucent and near-transparent sheets when
the scattering is in the forward direction. The term used is haze, defined
in ASTM Standard D 1003 as the light flux deviating from the incident
transmitted beam direction by more than 2.5. Hazemeters were devel-
oped and are commercially available for making the measurement of
haze and luminous transmittance. A conventional spectrophotometer
with an integrating sphere of the type illustrated in figure 4.22 may also
be used. The use of the term haze in this sense seems to be unique to the
plastics industry. In the paint industry, haze generally refers to the slight
scattering from a glossy surface that decreases distinctness-of-image gloss
(FSCT 1995). Other methods used in the plastics industry to measure
transparency and translucency are ASTM Standards D 1494 and D 1746.
The techniques for measuring and adjusting the opacity of poly(vinyl
chloride) sheets were discussed by Osmer ( 1978 ).
Charts for determining opacity
The paper industry was one of the first to utilize the Kubelka-Munk
equation for determining opacity. The original Kubeika-Munk equation
for incomplete hiding had been used since the late 1930s for determining
opacity in terms of the contrast ratio R
0
/Rw, based on precalculated charts
for a substrate Rw = 0.89. In its original 1931 form, the Kubelka-Munk
equation, shown as equation 4.25 below, could not be used to determine
the scattering coefficient directly.
(
R - R )! R -R (R -1/ R )eSX(J/R=-R=)
g = = = g =
R = ----------------
Rg -R= -(Rg -1/ R= )eSX(J/R= -R=)
(4.25)
K is not in the equation but is implicit in the relationship: K/S = (1 -
Rx)
2
/2Rx. Tables of the exponential function, eX, were used to compute
some theoretical information that could then be plotted for use in mak-
ing graphical solutions for specific situations (] udd 19 3 7). Specifically,
graphs were developed for the case where the reflectance of the white
substrate was 0.89, as used by the paper industry; or 0.80, as used by the
paint industry; or 0. 70, as used by the dental industry. The ordinate of
these graphs was the reflectance over black, R
0
; the abscissa was the hid-
ing power in terms of the contrast ratio, R
0
/Rw. Plotted on the graph
were the Rx and SX (to be interpreted as SCX to account for the concen-
tration as well as the thickness ). Thus, if one knew any two of these val-
ues, the others could be determined from one of the graphs for one of the
specific white substrates, Rw.
Use of such charts has now been largely replaced by comput-
ing techniques, easily carried out with a computer or programmable cal-
culator. The computing task was made straightforward when Kubeika
(1948) published the hyperbolic solutions of the original Kubeika-Munk
equation (eq. 4.25) described above.
Figure 4.23
Diagram showing the interrelation of
reflectance at incomplete hiding over
black, R
0
; reflectance at complete hiding,
Roo; TAPPI opacity, C
089
= R
0
/R
089
; and
SX, the scattering power times the thick-
ness. It must be remembered that a
change in concentration has the same
effect on the color, when hiding is incom-
plete, as a change in thickness, so SX is, in
reality, SCX; increasing the pigment vol-
ume concentration has the same effect
as increasing the thickness by the same
multiple (after Judd 1937).
Colorant Mixture 125
Why mention the old charts? Because a ready estimate of R,,
and SX (that is, SCX) can be obtained with a very simple tristimulus col-
orimeter, without any calculations at all, by making the Y, or G, filter
measurements of a sample over white and over black, as well as measure-
ment of the white substrate. Such a substitution of Y or G for R should
be used only for white or neutral materials. The chart is a quick guide.
Information for determining thickness or concentration corrections to
attain desired properties can be estimated from the charts without exten-
sive calculations. The chart for the white substrate of 89% (as used in
the paper industry) is reproduced here, as figure 4.23. The 89% reflec-
tance substrate is a good average for white pigments. (See the test meth-
ods published by the Technical Association of the Pulp and Paper
Industry [TAPPI 1992b, 1992d].)
Hiding power
Examination of the charts shows that when the contrast ratio-Y
0
/Y w or
R
0
/Rw-approaches 1.000 (100%), it becomes increasingly difficult to
read the graphs accurately. As a consequence, when these graphs were
0.90
0.90
0.80
0.80
0.70
0.70
0.60
0.60
0.50
0.50
0.50 0.70 0.80 0.90 1.00
TAPPI opacity, C0_89 = R01Raa9
126 Chapter 4
first used early in the history of determinations of hiding power, many
industries established a contrast ratio (
0
/Y wl of 0.98 as equivalent to
complete visual hiding. This is seldom strictly true, but that ratio is still
widely used as a standard. See ASTM Standard D 2805 for its use in the
paint industry.
The contrast ratio of 0.98 as complete hiding is strictly val id
only for perfect whites. Therefore, a contrast ratio of 0.98 does not
always correspond to complete vis ual hiding. Nonetheless, for many
applications, determination of satisfactory hiding power in terms of a
contrast ratio of 0.98, which is generally very close to complete visual
hiding, is acceptable. If an effort is made to increase the hiding, even to a
contrast ratio of 0.985, reference to the charts shows that the concentra-
tion of scattering pigment must be increased by at least 10-20%, adding
significantly to the cost and possibly altering physical properties of the
pigmented material as well. In many commercial applications such an
effort is deemed not worth the increased cost and formulation problems.
Incidentally, the small addition of an absorbing pigment such as black
will also improve the hiding, but, of course, this will lower the
reflectance, Re,,.
In many applications, with many materials, the degree of
visual hiding is not the most important consideration. Whenever the sta-
bility of a material that might be exposed to light is paramount, the spec-
trophotometric hiding must be considered. Perfect spectrophotometric
hiding is defined as a contrast ratio of unity (l.000) at every wavelength,
as calculated from reflectance measurements made over a perfect white
(R = 100%) and a perfect black (R = 0%) in the near-infrared and near-
ultraviolet spectral regions as well as in the visible. Incomplete spec-
trophotometric hiding is the contrast ratio calculated from reflectance
measured at the wavelength (or selected wavelengths) at the reflectance
maximum over a white of defined reflectance and a black of near-zero
( < 1 % ) reflectance. In the discussion that follows, the contrast ratio
based on the Y tristimulus value is abbreviated (CR)y, and the contrast
ratio based on reflectance measurements at specific wavelengths is abbre-
viated (CR)R. In figure 4.24 the two types of contrast ratios are illus-
trated (Johnston-Feller 1986) for two samples of alizarin-Ti0
2
mixtures.
Sample B appears to be very close to complete visual hiding, (CR)y, that
is,
0
/Y w = 0.976 when Rw = 1.00. However, in the red region, the (CR)R
average is only 0.877. This is clearly evident from an examination of t he
spectral curves illustrated; the reflectance in the red region for the sample
measured over black is much lower than the reflectance measured over
whi te. Sample A, on the other hand, is at incomplete hiding both in
terms of the (CR)y = 0.923, and in terms of the reflectance in the red
region, (CR)R = 0.727. The point is made in the cited 1986 article that
the rate of fading of sample B is 30% higher when exposed over white
and 50% higher when exposed over black, when spectrophotometric hid-
ing is incomplete in the red region, than when hiding is complete at all
wavelengths. The increased rate of fading over black is probably a heat
effect; the painted panel, which transmits in the red region, gets hotter
due to the black background's absorption of the infrared wavelengths.
Figure 4.24
Two formulations of alizarin lake mixed
with rutile Ti0
2
in poly(vinyl acetate)
(Vinac B-7) applied at incomplete hiding
over a white of 100% reflectance and a
black of 0% reflectance. The contrast ratio
based on the luminous reflectance, Y, is
designated as (CR) y- The contrast ratio for
the average reflectance in the red region
of 660-700 nm is (CR)w Note that (CR) y
for sample B is very close to the 98% con-
trast ratio widely considered to be equiva-
lent to complete hiding, but in the red
region of the spectrum, hiding is far from
complete. Sample A is lower in hiding, as
indicated by the lower contrast ratios.
Colorant Mixture 127
y 0 i.---- R ----
90
- measured over white
- - - measured over black
80 (CR)R = average CR at
660-700 nm
70
<ii 60
------
g
QJ
(CR)y = 97.6%
u
c
.s
u
50
QJ
;:;:::
1':'
.....
c
QJ

sample B Ti02 PVC: 11 %
QJ
40 a..
30
(CR)y=92.3%
20
10
sample A Ti0
2
PVC: 2.5%
400 450 500 550 600 650 700
Wavelength (nm)
When hiding is incomplete at all wavelengths, the rate of fading is more
than twice that for samples at complete spectrophotometric hiding at all
wavelengths.
The contrast ratios illustrated in figure 4.24 were calculated
for a white substrate reflecting 100 % . If the white substrate reflectance
were significantly lower than this value-for example, 80% (the average
reflectance of the white portion of the black-and-white contrast paper
widely used in the paint industry)-the measured contrast ratios would
be higher; from an analytical point of view, this would be misleading.
For sample Bin figure 4.24, the (CR)y would be above 0.99, compared
with 0.976 over the 100% reflecting white; the (CR)R would be 0.99,
compared with 0.877. The result would imply that the sample is very
close to complete hiding, possibly within experimental error, as it would
visually appear to be over a substrate of 80% reflectance. For sample A,
the spectrophotometric (CR)R in the red region would rise from about
128 Chapter 4
0.73 for a substrate reflectance of 100% to 0.95 for a substrate of 80 %
reflectance! Thus, when spectrophotometric, rather than visual, hiding
may be of prime importance in determining the effect of light exposure
on the performance of a specific thickness of a material or of a particular
colorant formulation, the reflectance contrast ratios at wavelengths of
maximum reflectance (minimum absorption) should be calculated for the
substrates of ideal reflectances: a white of 100% reflectance and a black
of 0% reflectance. Alternatively, the reflectance, Rw, of the intended sub-
strate-if known-can be used in the measurements and calculations.
With computers, the calculation to Rw = 100% is relatively straightfor-
ward. However, the concept is not always appreciated, and the calcula-
tion is not often done.
The reader may well question when a substrate of 100%
reflectance is ever encountered. The answer is that when the substrate is
a metal (a nondielectric), the specular reflection may be extremely high at
the specular angles, approaching mirror reflection. It is important also to
realize that the concepts described apply equally to the nonvisible por-
tions of the solar spectrum: the actinic ultraviolet regions as well as the
thermal infrared regions.
The list of materials requiring knowledge of translucency
and opacity is almost limitless. It includes paints, plastics, ceramics, and
textiles, as well as liquids, both natural and manufactured. Even the
scattering in gases-such as in clouds, which contain finely dispersed
particulates (water droplets, ice crystals, dust, etc.)-has been studied
extensively by use of radiation transfer theories such as those of Mie and
Kubeika-Munk.
Tinting strength: absorption and scattering
Of major importance to the user of colorants is their strength, that is,
their ability to alter the color in a particular manner when mixed with
one or more other dyes or pigments in a particular application. The term
tinting strength is generally applied to this characteristic. As most com-
monl y used, it is not an absolute measurement but a relative one; that is,
it is used in a comparative sense: one particular colorant may be found to
have a higher tinting strength than another when different colorant types
or different samples of the same t ype are compared. Tinting strength is an
important characteristic when determining relative costs of different col-
orants, when formulating or reformulating to achieve optimum character-
istics, or when ensuring quality control in purchasing and manufacturing.
As an example, an artist, when purchasing a tube of a partic-
ular colorant, may wish to know which brands are the best value. Price
alone may not be the best indicator. If a more expensive tube contains
more pigment per unit volume and thus has a higher tinting strength, it
may be the cheaper in terms of effect.
In the case of dyes, which do not scatter light, the absorption
values determine the tinting strength. Their strength is best determined
from a measurement of the absorption at the wavelength of the absorp-
tion maximum. This may be done by measuring the transmittance of a
solution of the dye and using the Beer-Bouguer equation; or by making a
Colorant Mixture
129
dyeing using a standard procedure, measuring the reflectance, and apply-
ing the single-constant Kubeika-Munk equation described earlier in this
chapter. The Beer-Bouguer equation for the absorption determined for a
transparent solution was originally introduced as equation 4.5c. It can be
restated as
or (4.26)
where Tis the decimal fractional transmittance, C is the concentration,
X is the cell thickness, and A is the unit absorption coefficient.
The relative tinting strength (TS) is calculated at the absorption-
maximum region of the sample (Spl) in question relative to the same mea-
sured absorption of the standard (Std) (Kuehni 1983) by the following:
%TS= 100 [ Aspl l
A std
(4.27)
When standard dyeings are being compared, the single-
constant Kubeika-Munk equation used on opaque materials relates the
reflectance to the absorption as
(l-R)2 = C ~ ]
2R S
or (4.28)
where R is the decimal fractional reflectance, C is the decimal fractional
concentration of the dyestuff, S is the substrate scattering, and K is the
unit absorption coefficient. The dyeing is made on identical standard sub-
strates for the sample and the standard; S is set equal to unity (1.000) for
the standard substrate. The relative tinting strength (TS) is then calcu-
lated in the same way (Garland 1983) with the following:
%(TS)K = 100 [(K I S)spl l
(K I S)srd
(4.29)
where (TS) K refers to absorption tinting strength. Recall that the Beer-
Bouguer absorption coefficient A is not the same as the Kubeika- Munk
absorpti on coefficient K. The Kubeika-Munk unit K is twice the value of
the Beer-Bouguer unit A (Judd and Wyszecki 1975).
The relative tinting strength of a dye determined by the two
methods-by measurement in solution and the use of equations 4.26 and
4.27, or by measurement on dyed fabric and the use of equations 4.28
and 4.29- may not be in agreement. The difference may be attributable
to variations in dyeability caused by the effects of processing and by pos-
sible interactions with other constituents used in the application.
In the case of pigments, the measurement of tinting strength
is generally based on the preparation of opaque samples-for example,
paint fi lms at complete hiding or plastic sampl es of sufficient thickness
to be opaque.
In chapter 3 it was pointed out that pigments may be placed
into one of three categories. Type I pigments scatter much more light
130 Chapter 4
than they absorb; an example is white pigments. Type II pigments absorb
much more light than they scatter; these include most blacks and many
transparent, primarily organic, pigments of small particle size. Type III
pigments both absorb and scatter light significantly; examples are many
inorganic pigments of higher refractive index and organic pigments, both
of larger particle size.
The tinting strength of type I pigments, such as whites,
depends primarily on their scattering coefficient, S. This is generally
determined by preparing a sample of the material mixed with a pre-
scribed amount of a standard reference absorbing pigment, such as a
black. For type I scattering pigments, the scattering strength is deter-
mined as follows:
%(TS)s = 100 [(K I S)srd l
(KI S)spl
(4.30)
where (TS)
5
refers to the scattering tinting strength. Note that in deter-
mining the relative scattering strength, the KIS ratios are inverted from
what they are in equation 4.29 for determining absorption tinting
strength. This is because S is in the denominator of the K/S ratio.
The calculation of the absorption tinting strength for absorb-
ing, low-scattering, type II pigments is described by equations 4.28 and
4.29, except that Sis based on a standard white pigment used in admix-
ture. The strength of type II pigments depends on their absorption
coefficient, K, and is generally determined by preparing a sample of a pre-
scribed amount of pigment in a mixture with a prescribed amount of the
standard reference white pigment. In this case, the form of the single-
constant Kubeika-Munk equation shown in equation 4.7 is generally used,
and K/S is determined at the wavelength of the absorption maximum. The
relative absorption tinting is then calculated by use of equation 4.29.
Determining the tinting strengths of pigments that both
absorb and scatter significantly-that is, type III pigments-presents a
more complex problem. The strength of these pigments depends on the
way they are to be used. If they are used to make a light color, for
example, a color equal to or above a Munsell Value of about 5 (20%
luminous reflectance), the white contributes almost all of the scattering,
and the type III pigment contributes the significant absorption. Such a
pigment's tinting strength is then attributable to its absorption character-
istics and may be calculated from equation 4.29, the same as for type II
pigments and for textile dyeings. In this case, the pigment's scattering
characteristics are not of great significance.
If a type III pigment is used in a mixture containing a small
amount, or none at all, of scattering pigment, the problem of its strength
becomes less clear. Is the strength dependent on its absorption or on its
scattering, or on both characteristics? An example of such a mixture is
the dark olive drab green that was once used to paint some telephone
company trucks. Until industry was required to use lead-free pigments,
this color was often made by mixing a yellow lead chromate pigment and
a black pigment; the black supplied the absorption, and the yellow sup-
Colorant Mixture 131
plied the scattering. A green color resulted because the chrome yellow
exhibited its maximum scattering in the green region. The absorption of
the chrome yellow did not matter; the black supplied the absorption. The
tinting strength of the yellow pigment in this application was due to its
scattering characteristics. Its tinting strength could be calculated from
measurement of a mixture with a standard black pigment according to
equation 4.30, as is used for type I white pigments. In this case of
wavelength-selective scattering, the scattering tinting strength would
be measured at the wavelength of the scattering maximum.
When a type III colorant, which both absorbs and scatters
significantly, is applied at incomplete hiding, as a glaze, over a reflecting
substrate such as white, its absorption primarily determines its color
because the transmitted light is subsequently reflected. The white sub-
strate provides the scattering. If the same colorant is applied at the same
concentration and thickness over an absorbing substrate such as black,
however, its scattering characteristics determine the color. The transmit-
ted light is absorbed by the black substrate. Thus, the yellow lead chro-
mate pigment described above, which was used to make olive drab colors
in mixture with black, also makes a low-chroma olive green when
applied at incomplete hiding over black.
Figure 4.14 presented earlier shows the reflectance curves of a
yellow lead chromate applied at incomplete hiding over a white substrate
and over a black substrate (Johnston and Feller 1967; Johnston 1973).
The maximum reflectance of the sample over black occurs at about 520
nm, in the green region. By contrast, a type II yellow-flavanthrone,
which absorbs far more light than it scatters-applied over black makes
a very small change in the black substrate color, but that small change is
to a slight green tinge, as illustrated in figure 4.15. Applied over a white
substrate, Flavanthrone Yellow results in a high-chroma yellow because
of its high yellow transmittance. Note that the glazes applied over black
are plotted at 5 times their actual values; the reflectance of the more
transparent Flavanthrone Yellow applied over black is less than 3% at
the scattering maximum (535 nm), whereas that of the chrome yellow
applied over black is close to 10%. These values are relative only.
In figures 4.16 and 4.17, two red pigments are illustrated as
applied over white and over black: cadmium red, a type III colorant that
scatters and absorbs significantly, and a more transparent type II pigment,
Quinacridone Red, which absorbs strongly but scatters little (Johnston
and Feller 1967; Johnston 1973). The reflectance of the Quinacridone
Red glaze over black is less than 1 % .
Some mixtures of type II and type III pigments result in seem-
ingly surprising colors. For example, years ago many commercial high-
chroma red paints and plastics were made by mixing a transparent type
II absorbing organic violet, or bluish red, in mixture with molybdate
orange, a type III pigment. That a very dark violet in mixture with a
bright orange would result in a bright red may, at first, seem surprising.
But the explanation is straightforward in terms of their absorption and
scattering characteristics. The molybdate orange absorbs most of the
light from 400 to 550 nm, thus absorbing the violet reflectance. The
132
Figure 4.25
Reflectance curves of molybdate orange
masstone, Quinacridone Red masstone,
and a mixture of 25% Quinacridone Red
with 75% molybdate orange. The mixture
(solid line) is a high-chroma red. In this
case the greater absorption of the
Quinacridone Red in the middle of the
spectrum shifts the curve so that the
reflectances in the yellow and orange
wavelengths are decreased, leaving only
the red wavelengths to be reflected.
The molybdate orange absorbs all of
the short wavelengths, where the bluish
Quinacridone Red has significant
reflectance in mixture with a white
pigment (dotted curve).
Chapter 4
violet, on the other hand, has its maximum absorption in the yellow and
yellow-orange region, 560 to 620 nm, thus removing these wavelengths
from the molybdate reflectance. The molybdate reflects highly in the red
region because of its high scattering, and the absorption of the violet
decreases in this long-wavelength region. The nonabsorbed reflectances
remaining are in the red region, so the color of the mixture is bright red.
In the case of a bluish transparent red mixed with molybdate
orange, the resulting red is of higher chroma than the mixture with a
violet. Figure 4.25 illustrates the curves of the masstone of molybdate
orange, type III, and of Quinacridone Red, a type II organic bluish red
pigment, as well as the curve of a mixture of 75% molybdate orange and
25% Quinacridone Red. In order to see the typical absorption bands of
the Quinacridone Red, obscured in the very dark masstone curve, the
curve of a mixture of 10% Quinacridone Red in Ti0
2
is included in the
figure. As can be seen by the curve of the mixture of Quinacridone Red
and molybdate orange, the color is a bright red.
90
80
70
60
50
40
30
20
10
y 0 -----R----
molybdate orange and Quinacridone Red mixtures
100% orange
100% red
10% red in Ti0
2
--- 25% red & 75% orange
..
.
.
.
...............
,,,,.--
.
.
.
/ .
.
.
.
. /
. .
: I
! i
! I
..
: I
.
: I
~ . '
. . I
: .. :
:' ..... f i
---
I I
\\ f I
... f i
. . .
. ..: I
.. .
.. . . . .. : ,.
.. . . ...
.
/
_______ _,,,,,
400 450 500 550
Wavelength (nm)
600 650 700
Colorant Mixture 133
Thus, we see that when a type III pigment, which both
absorbs and scatters significantly, is used in mixture with or applied over
a type I pigment such as white, which scatters far more than it absorbs,
the tinting strength of the type III pigment is dependent on its absorption
characteristic as calculated from its Kubeika-Munk K/S value. However,
if a type III pigment is used in mixture with or applied over a type II pig-
ment such as black, which absorbs far more significantly than it scatters,
it is the scattering characteristic of the type III pigment-as calculated
from its Kubeika-Munk S coefficient-that determines its tinting strength.
Nonetheless, in some two-pigment mixtures of types II and III, both the
absorption and the scattering characteristics of the type III pigment are
important, as exemplified by the molybdate orange in the Quinacridone
Red-molybdate orange mixture discussed above.
When both absorption and scattering properties are impor-
tant, each may be checked. Make a mixture with white and use equations
4.28 and 4.29 to calculate K's; then make a mixture with black to calcu-
late the S's using equations 4.28 and 4.30. A quick visual comparison
between two pigments can be made by applying each at incomplete hid-
ing over black and white substrates. If they are made at the same concen-
tration and thickness, their relative absorption strength can be seen from
the area coated over white, and their relative scattering strength from the
area over black.
In the case of mixtures of two or more type III pigments,
which both absorb and scatter light significantly, it is difficult to predict
which optical characteristic of each is of primary importance in deter-
mining the color of the mixture. Generally speaking, the only way to
determine this is by trying various amounts of each pigment in the mix-
ture. If a computer color-matching program is available, this can be done
without making up samples; if not, trials can be made by varying the
concentration of each pigment and observing what happens.
One of the reasons it is so difficult to predict which pigment
is primarily responsible for either the scattering or the absorption of a
type III pigment mixture is that typical type III pigments are inorganics
with refractive indices of 2.0 or above. The most frequent variation
encountered in these pigments is in their particle size. And the particle
size is dependent not only on how the pigment is manufactured but also
on how it is dispersed.
(Anyone who examines the works of old masters and com-
pares them with the works of modern artists knows that in the days long
before commercial dispersions were developed, artists used whatever pig-
ments were available from natural sources or from crude manufacturing
operations. They ground the pigments by hand, using a muller or just a
spatula; the result was paints containing pigments that were coarsely
ground in comparison with modern pigments produced commercially.)
Fortunately, changes in the particle size of type III inorganic
pigments with refractive indices of 2.0 or more affect the absorption and
scattering coefficients in a similar manner-that is, if the particle size is
either increased or decreased from the size required for maximum
absorption and scattering power, both absorption and scattering will
134
Figure 4.26
The change in the absorption coefficient ,
KA, and the scattering coefficient, Ks, as
the particle diameter of iron oxide pig-
ments is changed (calculated from Mie
theory [after Brockes 1964; Buttignol
1968]).
Chapter 4
decrease. The rates of change in absorption and scattering efficiency are
not the same, but they are also not so different that, qualitatively, the
concept fails to be a helpful guide.
Figure 4.26 is an illustration of the relationship of the
absorption and scattering coefficients to the particle size of iron oxide
pigments, as calculated from Mie theory (Brockes 1964; Buttignol 1968) .
In the previous section it was stated that Mie theory applied to single
scattering only. However, Mie theory also provides an excellent guide to
relative absorption and scattering ratios in multiple-scattering situations.
In the article by Buttignol, the qualitative agreement between the absorp-
tion and scattering coefficients for iron oxide pigments was demon-
strated, as determined both from Mie theory and from Kubeika-Munk
theory. From the graph shown in figure 4.26, it can be seen that the opti-
mum particle size for maximum scattering is virtually the same as that
for maximum absorption, and that either decreasing or increasing the
particle size from this maximum results in a decrease in both the absorp-
tion and scattering efficiency.
Although the curves in figure 4.26 were calculated for iron
oxide pigments, they are qualitatively applicable to other type III inor-
ganic pigments of refractive index greater than 2.0. For example, molyb-
20
10
6
4
2
.6
.4
2
.1
.01 .02
I
I
I
I
I
I
I
I
Ks I
I
iron oxide pigments
/'
I \
I \
I \
I \
I \
I \
I \
I \ /'
I \
I \
I \
I \
I ',
.04 .06 .1 .2 .4 .6
Particle diameter (m)
'
'
'
\
2 4
Colorant Mixture 135
date orange (fig. A.28) can be easi ly overground, that is, ground smaller
than the optimum particle size in the dispersion process. When this hap-
pens, both the absorption and scattering coefficients are decreased, as
would be predicted from the curves in figure 4.26. The masstone color
becomes dirty because of the decrease in the orange and red scattering,
and a mixture with white pigment, such as Ti0
2
, shows the overground
pigment to be paler as well, due to the reduction of the absorption.
It is not possible to categorize absorbing pigments as either
type II or type III solely on the basis of refractive index, because organic
pigments of refractive index of 2.0 or less can both absorb and scatter
significantly if the particle size is in the proper range to achieve scattering.
In figure 4.27, Brockes's curve for organic pigments, calcu-
lated from Mie theory, is shown (Brockes 1964). It can be seen that the
shapes of the absorption and scattering curves are dependent on the par-
ticle sizes as well as on refractive indices. They are very different from
the curves for the iron oxide pigments and other pigments of refractive
index greater than 2.0 illustrated in figure 4.26. Also apparent from
this graph is that the greatest absorption is reached (absorption tinting
strength) at the smallest particle sizes. (The calculations were stopped
before the absorption increased as particle size decreased further. )
Organic pigments (toners) do not lose absorption strength as the particle
size is decreased as do the type III pigments. In the past-in general,
prior to the 1970s-organic pigments were almost always manufactured
so that the scattering was low and the absorption high-typical type II
pigments. Such organic pigments, as manufactured to achieve maximum
absorption tinting strength, have little hiding power because their scatter-
ing is so low.
Prior to the 1970s, many of the bright yellows, oranges, and
reds in paints and plastics were pigmented with colorants containing
heavy metals, such as lead, which is dangerously toxic to animals and
humans if ingested, because the metallic elements accumulate in the
body. Other heavy metals, such as cadmium, used in plastics were less
commonly encountered but were also unhealthy if ingested. Soon there-
after, a great movement in pigmentation changes occurred to replace the
materials made with pigments containing heavy metals, particularly lead,
which had been used for so many years in the form of lead white and
also as lead chromates for making bright yellows and oranges. Many of
the organic pigments introduced as replacements are completely metabo-
lized by the human body and represent no problem if ingested.
Examination of the Mie curve (fig. 4.27) shows that if one is
willing to sacrifice absorption strength, one can make an organic pigment
that scatters significantly by increasing the particle size. As a result of the
mandate to eliminate lead from commercial paints and plastics, changes
have come about in industrial pigment production so that organic col-
orants with larger particle size and lower absorption tinting strength but
higher scattering strength may be encountered.
Schafer and Wallisch ( 1981) described this trade-off in optical
properties for Pigment Yellow 74 (Dalamar Yellow), a Hansa-type
organic pigment. They illustrated the curves of pigment particle size
136
Figure 4.27
The change in the absorption coefficient,
KA' and the scattering coefficient, Ks, as
the particle size of organic pigments of
refractive index less than 2 .0 is changed
(calculated from Mie theory [after Brockes
1964])
Table 4.3
Relative absorption tinting strength and
opacity of samples of Pigment Yellow 74
(PY 74) compared to their average particle
size (Schafer and Wallisch 1981 ).
Chapter 4
20
10
6
4
2
.6
.4
.2
.1
.01 .02
I
I
I
I
I
I
I
Ks/
I
.04 .06
organic pigments
,.---
/ ......
/ ',
I
I
I
I
I
.1 .2 .4 .6 2 4
Particle diameter (m)
distribution and also provided electron micrographs for three different
versions of PY 74: one with high (absorption) strength; a standard; and
another with high opacity (high scattering). The median particle sizes
were 0.15 m for the high-absorption-strength type; 0.22 m for the
standard; and 0.33 m for the high-opacity version. The relative effects
on the absorption tinting strengths and the opacities (indicative of the
relative scattering tinting strengths) for these pigments, as used in a long-
oil alkyd paint, are summarized in table 4.3. The trade-off between
absorption tinting strength and scattering strength (described as relative
opacity) illustrated in this table is in qualitative agreement with the
Mie curve for organic pigments illustrated in figure 4.2 7. An excellent
overview containing illustrations of y-quinacridone pigments of different
particle sizes, presented in electron micrographs as well as in color
plates, is that by Gerstner (1966).
Median particle Relative tinting
Type size (m) strength Relative opacity
High-strength PY 74 0.15 120 42
Standard PY 74 0.22 100 100
High-opacity PY 74 0.33 58 175
Figure 4.28
Spectral reflectance curves of alizarin (1 ,2-
dihydroxyanthraquinone) mixed with Ti0
2
.
The pure alizarin is yellow; the calcium-
alumina lake is the familiar bluish red.
Colorant Mixture 137
The type II pigments illustrated by the Mie curve of figure
4.27 may also be types that contain inorganic constituents. Some are
lakes, which are dyes precipitated on inorganic bases or substrates. An
example is the calcium-alumina lake of alizarin. The organic chemical,
in pure form, is a yellow, but when precipitated on inorganic aluminum
hydrate, it forms a brilliant, unique red. The spectrophotometric reflec-
tance curve for the pure alizarin ( 1,2-dihydroxyanthraquinone) dispersed
in a rutile Ti0
2
paint system is shown in figure 4.28. The spectral curve
is obviously that of a yellow of moderate chroma. Also included in this
figure is the spectral curve of the calcium-alumina lake of the chemical
dispersed in the same rutile Ti0
2
paint system. It is readily apparent
that the lake is the result of a reaction between the 1,2-dihydroxyanthra-
quinone and the calcium-alumina carrier. Not all transparent organic
chemicals, when made into a lake on inorganic particles, change hue so
radically. Many of the modern transparent organic pigments are offered
for sale either as the lake or as the pure material, called a toner.
90
80
70
]i
60
g
<IJ
u
c
<U
ti
50
<IJ
""

c
<IJ
u
Qj
40 Q._
/
/
30
___ ,
20
10
400 450
y 0 -----R----
I
/
/
I
I
I
mixtures in rutile Ti0
2
I
I
I
I
I
I
I
/
/
/
/
/
--
.,,.
.,,.
-- alizarin lake
- - pure alizarin
500 550 600 650
Wavelength (nm)
--
700
138
Figure 4.29
Universal scattering curve for nonabsorb-
ing pigments as a function of particle
diameter, D (in m) ; wavelength, le
(in m); and the ratio, m, of the refrac-
tive indices of the pigment, n P' and the
vehicle, n v (after LaMer et al. 1945).
...,
c
4
C 3
i:
Q)
0
u
b.O
c
~
u 2
Vl
Chapter 4
One of the type II pigments of important historical interest
is iron blue, also called Prussian blue. Because the word "iron" is in its
name, it may be confused with an inorganic pigment, but this is a trans-
parent, complex organic (carbon-containing) pigment-ferric ferro-
cyanide-which has a low refractive index (n = 1.56 at 560 nm) and a
high tinctorial (a bsorption) strength.
In the case of Ti0
2
, a type I pigment that does not absorb
appreciably, the Mie scattering can be predicted from a simple graph.
It has been called the universal scattering curve and was first published
by LaMer (LaMer et al. 1945) and later by DeVore and Pfund (1947). It
is reproduced here as figure 4.29. The abscissa contains the values for
nv, the refractive index of the vehicle; nP, the refractive index of the
pigment; D, the diameter of the white pigment particles expressed in
micrometers (m); A, the wavelength (m); and m, the ratio nP/nv. Mitton
(1973) shows that the opacity and tinting strength of white pigments is
linearly proportional to the value of M
2
, when M = (n/n
0
)
2
- l/(n/n
0
)
2
+ 2, where n
0
is the refractive index of the vehicle and nP is the index of
the pigment. An approximate expression for M is M ""0.4(nP - n
0
) . In
Mitton's figure 2, he relates the square of M for ruti le and anatase Ti0
2
,
lead white, ZnO, calcium carbonate, and BaS0
4
to their relative hiding
power (or scattering tinting strength).
Gerstner (1966) presents electron micrographs and color
plates of two Ti0
2
pigments of different particle size. The color plates
show the two whites tinted with blue to illustrate the differences in the
tinting strengths of the two white pigments. The Ti0
2
, which is -0.25 m
'\
'
'
'
universal scattering curve
( LaMer)
--
,,. ....
/ ....
' /
.... -......
....
....
....
0 ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ - - ' - ~ ~ ~ ~ - ' - ~ ~ ~ ~ - J . . . ~ ~ ~ ~ - - ~ ~ ~ ~ - - ' - ~ ~ ~ - - - - '
0 0.2 0.4 0.6 0.8 1.0
n v(Ol/..)(m2 - 1 )/(m2 + 2)
Where m = nplnv
1.2 1.4 1.6 1.8
Colorant Mixture 139
in diameter, has more lightening effect-that is, scattering power-than
the Ti0
2
at - 0.38 m.
Throughout these remarks concerning the tinting strength of
colorants, details of measurement techniques have thus far not been pre-
sented. If color-measuring equipment is not available, two samples may
be tested visually. Paints of both are prepared identically and then com-
pared visually. To obtain a quantitative measure, a second preparation of
the sample in question can be made at a concentration estimated to make
the two match. If the first adjustment is not correct so that the color of
the second preparation concentration does not visually match the refer-
ence standard, the procedure is repeated. To obtain an exact estimate
by this trial-and-error procedure is time consuming, but it can be done.
Often an absolutely exact, quantitative measure is not necessary; an
approximation will suffice.
If color-measuring instruments are available, relative tint-
ing strengths can be calculated with the appropriate equations (eqs.
4.26-4.30). It was stated earlier that the colorant absorption strength is
measured at the absorption maximum. Using a spectrophotometer, the
wavelength of the absorption maximum can readily be determined. To
determine the tinting strength, the measured reflectance is converted to
KIS and the relative strength calculated with equation 4.29 or 4.30.
But tinting strength can also be measured with a tristimulus
colorimeter, provided that the following procedures are followed: ( 1) If
the colorimeter reads directly in terms of R, G, and B (red, green, and
blue) filter measurements, the values are used directly as reflectances in
the appropriate calculations; no corrections are necessary. (2) If, how-
ever, the tristimulus colorimeter measures in terms of CIE tristimulus val-
ues, each tristimulus value must be normalized to unity, that is, to R, G,
and B reflectance. The tristimulus values obtained with a tristimulus filter
colorimeter are generally in terms of illuminant C and the 1931 2 stan-
dard observer. Thus, for the reference white, X = 98.0%, Y = 100.0%,
and Z = 118.0%. Because the X value contains a portion (20%) of the
blue reflectance, it must be corrected to obtain the red reflectance, R,
which would be obtained if a filter with only red transmission had been
used. The following equation group describes the conversion of the tri-
stimulus values, X, Y, and Z, to a reflectance basis, that is, in terms of R,
G, and B filter reflectances to be used in calculating the tinting strength:
z
B= -
1.18
R = _( X_I o_.9_8_-_o_.2_B_)
0.8
(4.31)
The values are all expressed as decimal fractions. The R, G, or B value
is converted to KIS, and the appropriate equation for tinting strength is
used: equation 4.29 for absorption strength and equation 4.30 for scat-
tering strength. To measure the scattering strength of a white pigment
from a mixture with black, the G reflectance is used.
In recent years colorimeters have become available that use
a dispersive element (such as a prism, grating, or interference wedge )
instead of filters. Known as spectrocolorimeters, they provide only colori-
metric data, such as tristimulus values, and not the spectral data from
140 Chapter 4
which the colorimetric data are derived by computation. From such spec-
trocolorimeters, tristimulus values for other illuminants (such as D65)
and for other observers (such as the 1964 10 standard observer) may be
obtained. In this case, other equations must be used to normalize the tri-
stimulus values. For example, the following group of equations may be
used to calculate the reflectances used in tinting strength calculations for
illuminant D65 and the 10 standard observer:
Z10
B=--
1.073
G =Y
10
( X
10
I 0.9481- 0.199B)
R = ---------
0.8
(4.32)
For calculating relative tinting strengths, the filter (X, Y, or
Z; or R, G, or B) to be used for the measurement is selected on the basis
of the hue of the colorant being tested. The filter reading is treated as if
it were a reflectance value and is converted to an absorption value such
as KIS, according to equation 4.28. This technique of using tristimulus
filter measurements can be thought of as akin to using an instrument
with broadband filters. For blue and green colorants, use the red, R, filter
measurement to calculate the relative tinting strength (their absorption
maxima are in the red reflectance region); for purple and red colorants,
use the green, G, filter measurement (their absorption maxima are in the
green region); and for yellow and orange colorants use the blue, B, fi lter
measurement (their absorption maxima are in the violet and blue region).
Details of this procedure for calculating relative tinting strengths with
tristimulus colorimeters are presented in ASTM Standard D 4838. The
method is based on the work of Johnston-Feller and Bailie (1982b). In
these publications it is pointed out that the best accuracy and repro-
ducibility are obtained if the samples are made up to have a reflectance
of 35% to 45% at the absorption maximum. As the reflectances deviate
from this range, either up or down, both visual and instrumental sen-
sitivity decreases.
Also emphasized in these articles is the advantage of using
such broadband filter measurements in certain instances, as opposed to
using higher-resolution spectrophotometric measurements at a single
wavelength. Basically, there are two advantages. The first is the case in
which the chroma of one colorant is higher than the chroma of another;
corrections can be made by use of the filter measurements made at the
region of maximum reflectance, or minimum absorption. If only the
region of maximum absorption is used, the presence of a contaminant,
such as dirt, cannot be easi ly detected, and the tinting strength would
be too high because of the extraneous absorption of the contaminant.
Likewise, when comparing relative tinting strengths of different pig-
ments, the chroma can be as important as the tinting strength. Hence,
the correction can be made by subtracting the absorptions at the region
of maximum reflectance. The technique for making such corrections is
described in the two previously cited publications.
The second advantage is in comparing colorants that have
slightly different specific absorption maxima-that is, different spec-
trophotometric curves; the integrated reflectances over a broad range of
Figure 4.30
Spectral reflectance curves of indanthrone
blue (PB 64) and ultramarine blue (PB 29)
in mixture with rutile Ti0
2
. Comparison of
the concentrations indicates that the tint-
ing strength of PB 64 is almost ten times
that of PB 29.
Colorant Mixture 141
wavelengths in the region of absorption are often more suitable than the
reflectance measured at a single wavelength. The latter situation occurs
when colorants of similar hue but different chemical composition (and,
as a result, slightly different spectral curves) are compared. Relative tint-
ing strength is generally considered when colorants of similar chemical
composition are compared. However, there may be times when the rela-
tive value of two different chemical types-widely different in price and
strength-may be of interest. An example is the comparison of the rela-
tive values of red-shade indanthrone blue (PB 64) and ultramarine blue
(PB 29), both reddish blues. (PB 29 and PB 64 are Colour Index names;
see appendix C.)
Indanthrone blue has much higher tinting strength than ultra-
marine blue, but it is also more expensive. The spectral reflectances for
two concentrations of each blue mixed with rutile Ti0
2
are illustrated in
figure 4.30. It can be noted that indanthrone blue has about ten times as
100
90
80
70
60
50
40
30
20
10
- V --11-f----- B ------- G
y 0 14----R----
ultramarine blue (PB 29) and indanthrone blue
(PB 64) mixtures in rutile Ti0
2
I \
I /"X .
. . .
i:: ' .....
.. : ' ..
I
. .. .
.. : ' .
: ., ..
. .
' ... . .
\ ...
. .
PB 29 50 : 50
-PB645: 95
-- PB 29 10 : 90
- - PB 64 1 .25 : 98. 75
....... _____ ,
\.... /.
,.... .
~ ./
'.. ./
.,.... ./
. ~ .......
........... __ . .,
....
..
.
.
.
.
.
400 450 500 550 600 650 700
Wavelength (nm)
142
Table 4.4
CIE notation for two blues: PB 29, ultra-
marine blue, and PB 64, indanthrone blue
(illuminant C, 1931 standard observer).
Sample Cone. %
PB 29
Ultramarine blue 50
PB 64
lndanthrone blue 5
PB 29
Ultramarine blue 10
PB 64
lndanthrone blue 1.25
Table 4.5
Examples of spectrophotometric and tri-
stimulus methods for calculating relative
absorption tinting strength (TS) by com-
paring PB 29, ultramari ne blue, and
PB 64, indanthrone blue.
Chapter 4
much tinting strength as ultramarine blue. Careful inspection of the con-
trast in the curves between the reflectance maximum and the reflectance
minimum would lead one to expect that the ultramarine blue is higher in
chroma (purity)-not important if high-chroma blues are not required.
The spectral curve shapes, however, are quite different, so it is difficult to
compare the relative tinting strengths of the two blues. In table 4.4 the
CIE descriptions of the two blues are summarized for the 1931 standard
observer and illuminant C.
The spectrophotometric data and the tristimulus data allow us
to compare the two methods for calculating tinting strength: (1) use of
spectral reflectance at the absorption maximum (reflectance minimum),
and (2) the tristimulus values from the red broadband filter, calculated in
this case from the X tristimulus value (or as measured directly on a tri-
stimulus colorimeter measuring R, G, B values). Using spectrophotometric
reflectance data, the analyst would select the wavelength of maximum
absorption for each of the blues: 630 nm for the ultramarine and 580 nm
for the indanthrone blue. He would compare the two for each of the pairs
closest in reflectance, using equation 4.28 for spectral reflectance data to
calculate unit K/S; equation 4.31 for calculating R from X; and equation
4.28 again, assuming the R filter value is reflectance. The relative tinting
strengths are then calculated with equation 4.29. Table 4.5 summarizes
the results of these calculations. Ultramarine blue (PB 29) is used as the
standard for these calculations. The overall average tinting strength is
9.45 0.5% for the two methods and two concentrations. As can be seen
from these data, the agreement between the two methods of determining
tinting strength is excellent, particularly considering the significant differ-
ences in the spectral reflectance curve shapes of the two blues.
Returning now to the original question-which blue repre-
sents the better value? From a recent price list (Kremer Pigments, Inc.
x
y
z x y Dom. /.. %P
25.26 24.16 63.82 0.2231 0.2134 474.5 44
24.37 21.93 56.65 0.2367 0.2130 470.0 40
44.98 45.25 85.38 0.2561 0.2577 476.0 26.5
41.58 39.88 77.22 0.2620 0.2513 471.0 25.6
Concentration
of PB 29 (%) Concentration Calculation method TS Average
Standard of PB 64 ( %) PB 29 PB 64 PB 64/PB 29 TS
50 5 630 nm 580 nm 9.6
10 1.25 630 nm 580 nm 9.2 9.4
50 5 X to R filter 9.7
10 1.25 X to R filter 9.3 9.5
Plate 1
Can you tell by looking at the exposed
samples in the lighter column (right) how
much alizarin red pigment was lost from
the corresponding samples in the darker
column (left)? We humans cannot tell by
visual examination the amount of pigment
lost by fading when comparing the faded
samples to the original unexposed samples.
However, the relative amount lost can be
calculated based on measurements of the
reflectances of the samples. The answer to
the question above is that the samples in
the faded column on the right have all lost
30% of the alizarin initially present, even
though the lightest and the darkest samples
in the faded column appear to show less
difference with their corresponding samples
in the unexposed column than do the other
sample pairs (from Johnston-Feller 1986,
with permission).
Plate 2
This photograph shows a three-dimensional
graph constructed to illustrate the depth of
fading in a series of paint films pigmented
with rutile Ti0
2
and alizarin and exposed
in a Fade-Ometer. The two variables illus-
trated are the concentration of the alizarin
relative to the total pigment concentration
(the scale on the x axis, increasing to the
left from the origin) and the pigment vol-
ume concentration (PVC) of the Ti0
2
(the
scale on the y axis, increasing toward the
right and front). The dowels are located
according to their initial pigment concen-
tration. The height of the dowels corre-
sponds to the depth of fading that was
measured; the color of the dowels reflects
the color of the unexposed samples; and
the color on the top of the dowels (white
or black) corresponds to the color of the
substrate. The precision was poor, but the
general tendencies are readily apparent:
increasing the concentrations of either pig-
ment decreases the depth of the fading.
This is what is expected. The depths of
fading were measured microscopically by
several observers who examined cross
sections of exposed samples imbedded
in beeswax, using a calibrated grid in the
eyepiece of the microscope. The observers
repeated the task several times. The aver-
age is illustrated.
+
t
-
'I
-
\
L
"
"
-- -.
., ,
--:1
.....
(3a) Unexposed (3b) 106 hours
,- -
,.,.. .,.
- '
t
-
t
\
L ~
"
t
,",,:," "1 .. ,
-'
,
.")'
r 1/
(
-
~
(3d) 286 hours (3e) 393 hours
(3c) 219 hours
-
-:-, 7- t'
(3f) 530 hours
Plates 3a-f
This sequence of images shows a portrait painted on a Mylar sheet;
five concentrations of alizarin in glazes were used to model the
sleeves. The painting was backed with white and mounted around
the drum of a Fade-Ometer in which it was exposed to the light
from a xenon lamp for a series of exposure times. The squares to
the left of each image illustrate the five different concentrations of
glaze used. After the painting had been in the Fade-Ometer for a
specified time, it was removed for measurement of the reflectance
and for calculation of the concentration of the alizarin remaining in
each of the five glazes. At that time, the portrait was also photo-
graphed. The resulting photographs are reproduced here with their
total exposure time in the Fade-Ometer noted: (a) unexposed
painting; (b) 106 h; (c) 219 h; (d) 286 h; (e) 393 h; and (f) 530 h,
which is roughly equivalent to a hundred years on a museum wall
well illuminated by diffuse daylight.
/1 7
/ 16
/ 15
'"
/ 14
E
e
..c
/13
u

C
:::J
/ 12

/ 11
/ 10
/ 9
/8
2
Plate 3g
3 4 5 6
Munsell Value
o Unexposed
Exposed
7 8
Thi s graph shows the changes in Munsell Chroma and Munsell
Value of the alizarin pigment in the painting's sleeves due to fading.
The Value increased on all samples as the alizarin faded. The initially
darker (low Munsell Value) glazes, numbered 1, 2, and 3 on the
graph, increased in Chroma as fading proceeded; the maximum
Chroma was reached by the middle concentration, glaze 3. Glazes 4
and 5, the lightest glazes, decreased in Chroma as the color faded.
It is this last type of color change that we customarily recognize as
fading. That darker colors actually increase in Chroma, or saturation,
thus appearing to get brighter as they get lighter, is often not under-
stood to be the result of fading. All chromatic colorants do this when
applied in high concentration at incomplete hiding over a reflecting
substrate.
Plates 5a, b
Photographs of the Aach print made with a
fluorescent orange pigment, as described in
chapter 6. The photo on the left (a) was
taken in daylight; the one on the right (b)
under BLB black light. When the photo-
graphs were taken, the print, framed under
glass, had been hanging on the wall of the
author's office for seventeen years, during
which time it was illuminated by diffuse
daylight and by ceiling fluorescent lumi-
naires. The fluorescence in the print has
remained strong, evidence that the artist's
special formulation of the fluorescent pig-
ment apparently did prolong its effective-
ness. Normally the pigment's light stability
is found to be poor, and its use by artists
has been questioned by scientists for
this reason.
Plate 4
Shown are two glass-fiber gel coats of
identical pigment composition but with
different surface structures. The panel on
the right has a high-gloss surface; the one
on the left a matte surface. The surface
reflection from the glossy panel is sharply
reflected at an angle away from the cam-
era's view. The reflection from the matte
surface travels at all angles and is therefore
included in the camera's view, lightening
the red color as though white pigment had
been added to the red pigment. When the
two panels are measured in such a manner
that the surface reflection is included , the
measurements are nearly identical. The
(5a)
spectral curves of these two panels made
with surface reflection excluded are shown
in figure 6.5. The glossy sample reflects
about 1 % of the light over half of the
visible spectrum, while the matte panel
reflects about 5%. This reflected light-
violet, blue, green, and yellow-is mixed
with the red light reflected by the red pig-
ment particles, and it lightens the red
color. That both panels really are painted
with the same pigment composition is illus-
trated when the total reflectance is mea-
sured-that is, the surface reflection plus
the red pigment reflection-as shown in
figure 6.7.
(5b)
Colorant Mixture 143
1992), red-shade indanthrone blue is about 8.2 times as expensive as
ultramarine blue ($25 .35 per kilogram for the ultramarine and $206.80
for the indanthrone blue). However, indanthrone blue is 9.4 times higher
in tinting strength, as shown in table 4.5. So the expensive organic pig-
ment is actually a little less expensive in terms of tinting strength. Other
characteristics may, however, dictate the choice between the two. The
ultramarine blue is higher in chroma (purity), that is, it is cleaner in mix-
ture with white (see table 4.4). However, ultramarine blue has poorer
hiding power, particularly in dark shades, than the indanthrone blue
when mixed with white to make similar dark blues.
These comments are not meant to promote one pigment over
the other but are given to illustrate the concepts and calculations pre-
sented. In this particular example, there is no clear decision between the
two on the basis of price alone.
Confusion concerning the identification of colorants used in
objects made in the past can result from the practice by some manufactur-
ers of toning their colorants with other colorants of different chemical
type to adjust them to a standard. For example, a chrome yellow might be
toned at one time with a little Hansa yellow and at another with a bit of
benzidine yellow, to adjust the hue and absorption strength. Fortunately,
this practice is much less common today. One must remember why this
practice came about. It is virtually impossible to manufacture pigments
each time with identical optical, solubility, and dispersibility characteris-
tics. Prior to the 1960s, the practice of toning with other colorants was
not uncommon. Today, however, most modern manufacturers can pre-
cisely select and blend different batches to meet a standard. Computer
techniques make this practical.
With the sensitive analytical technique of solution spectro-
photometry (Saltzman and Keay 1967), even very small amounts of
organic colorants may be identified. The compositions of watercolors
labeled "indigo," reported by Johnston and Feller (1963), were checked
not only by reflectance curve shape but also by solution spectrophoto-
metry. Reflectance curve shape alone could not have identified the com-
plex mixture found in the sample that produced curve 3 of figure 4.31,
which is taken from the Johnston and Feller article. When solution tech-
niques were used, phthalo blue, phthalo green, and pyrazolone red were
found in sample 3, in addition to ultramarine blue and black, which were
obvious from the reflectance curve. The three organic pigments found by
solution spectrophotometry were present in small amounts as determined
by the subsequent use of differential spectral curve analysis. Who added
these? Did the pigment manufacturer tone the ultramarine blue with a
little pyrazolone red? Or were all of these mixed together by the artists'
tube watercolor manufacturer? If one found such a mixture on a work of
art, how could one know where all these pigments came from? Perhaps
the artist mixed all of them? (If the brushes were not cleaned well, he
might have.) One must systematically look for clues in many other areas
of each different color on the object. One could spot phthalo blue and
phthalo green, as well as ultramarine blue, in various areas of a painting
by their reflectance curves; pyrazolone red also has characteristic double
144
Figure 4.31
Spectral reflectance curves of four tubes of
watercolors, all labeled "indigo." Only
curve 4 is the real indigo. Curve 1 is a mix-
ture of phthalocyanine blue and black;
curve 2 is iron blue, ultramarine blue, and
black; and curve 3 is ultramarine blue,
phthalocyanine blue, phthalocyanine
green, pyrazolone red, and black (after
Johnston and Feller 1963).
Chapter 4
y 0 -----R----
90
80
70

60
:J

:s
Q)
u
c
_:g
50
u
Q)
q::

c
Q)
u
Q:; 40
(l._
30
20
2
10
400 450 500 550 600 650 700
Wavelength (nm)
absorption maxima . (See chap. 8 for a description of a systematic proce-
dure for approaching this problem.)
One last word about tinting strength. If one were comparing
two artists' colors labeled "indigo" for relative tinting strength, and if one
of them were made with phthalo blue and black and the other with real
indigo, the phthalo blue would, if not diluted to the same strength as
indigo, be the stronger. But if one indeed wanted to use genuine indigo,
analysis is advisable. Many of the major art materials manufacturers are
honest today in their labeling, but the practice is not universal, and it cer-
tainly was not in the past. Even if an artist has left a record of the names
and brands of colorants used in a particular work of art, the specific pig-
ments involved may prove on identification to be quite different.
In summary, the relative tinting strengths of colorants
of similar hue may be determined by the techniques outlined above.
Nonetheless, tinting strength may not be the only characteristic desired.
Colorant Mixture 145
Comments on Colorant Calculations and Identifications
Included in this section are discussions of the following: ( 1) some aspects
of the use of the Kubeika-Munk theory for colorant concentration calcu-
lations and colorant identification; (2) the use of reflection density, which
is a measure of reflectance used primarily in the printing industry; and
( 3) special single-number color scales related to both reflecting and trans-
mitting materials.
Quantitative limitations in the use of the Kubeika-Munk equation
Many factors affect the precision achieved by use of the Kubeika-Munk
equation for calculating colorant concentrations. In general these factors
can be placed into one of two categories: limitations of the equation
itself, and imprecision in the preparation of the colored materials.
Three constrictions inherent in the equation have been
alluded to in the previous discussions: ( 1) The relationship of the absorp-
tion coefficient, K, and the scattering coefficient, S, to the reflectance,
R, applies only to a single wavelength. (2) No account is taken of the
reflectance at the interface between materials, such as air and the sur-
face. (3) The assumption is made that the scattering is isotropic. Not
specifically mentioned were the assumptions that the incident light as
well as the reflected light are both diffuse; that the pigment particles
are randomly oriented; and that they are uniformly dispersed within
the medium. In the real world, such idealizations are difficult to realize.
Attempts have been made to devise more complex equations that more
nearly describe the real-world situation. See, for example, papers by
Orchard (1968), Atkins and Billmeyer (1968), and Richards (1970).
The simple relationship described by the Kubeika-Munk
equation is extremely useful as a guide for measuring the variations in
color encountered in the preparation of colored products. It is extremely
difficult to prepare colored products so that each preparation of a certain
formulation is identical, exhibiting no visual color difference from the
standard. For example, small errors in weighing the colorants in a for-
mulation mixture can result in detectable color differences (see Billmeyer
and Phillips (1974] and Stanziola (1980]). Thus, balances used for
weighing each colorant must be appropriately accurate for the amount
specified. Otherwise, the prepared mixture may not correctly represent
the desired formulation.
Johnston (1969) listed twelve areas in paint mixture prepa-
ration where problems of poor practice may result in inadequate color
control. In addition to weighing errors, some of the sources of varia-
tion listed include (a) variations in colorants from batch to batch;
( b) variations in behavior of colorants when used in different media
and on various substrates; (c) variations in measurements; (d) varia-
tions in sample preparation; and (e) use of nonrepresentative samples.
One of the least recognized sources of poor color control involves
the surface characteristics of the samples. Even slight differences in
surface reflection, whether from a glossy surface or a matte (scattering)
surface, can affect the perceived color and can be difficult to identify.
146 Chapter 4
In chapter 6 the importance of surface reflection characteristics is dis-
cussed at greater length.
In an experimental study of the reproducibility of making a
very high quality paint, Johnston (1963) pointed out the care with which
measurements and visual evaluations must be made so that the closeness
of the sample color match to the standard could be ascertained with a
high level of confidence.
In view of these difficulties, the fact that use of the Kubelka-
Munk equation for calculating colorant concentrations seldom gives per-
fect results is not particularly critical in the practical world.
In the case of paints, this technique also allows for small dif-
ferences in the behavior of colorants in different vehicle types to be taken
into account: a paint company may sell fifty different paints, but only a
few basic calibrations for the major types need to be made. Adjustments
for paints formulated in similar types of vehicles can be calculated from
a single set of calibrations by use of predetermined behavior characteris-
tics based on carefully selected mixtures for two reasons: ( 1) A deviation
of the preparation from the calculated amounts, if calculated consis-
tently, becomes less significant on the second preparation; for example,
a 10% error on the first preparation becomes an error of 10% of 10%-
or 1 %-on the second. (2) A difference in tinting strength ascertained
on the first preparation can be allowed for on the second. For precise
work, however, colorant calibrations prepared over a range of concentra-
tions are desirable.
Another admonition given by Johnston (1969) is that,
when possible, flexibility should be provided when a calculation is
made to match a desired color-that is, there should be three degrees
of freedom, one for each direction in color space: lightness-darkness,
redness-greenness, and yellowness-blueness. In the case of pigments,
this requires the incorporation of four pigments. Significant metamerism
may be avoided by the judicious incorporation of small amounts of
minor tinting colorants to allow for small directional differences
(Johnston 1973). In the case of pigmented systems, it is wise to include
a little black to allow for dirtiness, and a little yellow to allow for vari-
ations in the yellowness of the vehicles. Likewise, in the case of dyes,
a little yellow in the formulation-which can be withheld initially and
added only if needed-allows for yellowness in the substrates when
applied to textiles or pa per.
The various curves for different concentrations of colorants
shown in appendix A were computed from a single set of absorption
and scattering values using two-constant Kubeika-Munk theory, incor-
porating surface corrections according to equation 4.12b. These calcula-
tions were made for the purpose of preparing the illustrations. If one
should attempt to calculate the concentrations shown using simple single-
constant Kubeika-Munk theory, the results would not exactl y reproduce
these curves because of the difference in the equations used, but they
would be good rough approximations.
Colorant Mixture 147
Qualitative applications of the Kubeika-Munk relationship
One of the major points of emphasis throughout this monograph has been
the identification of individual colorants from their spectrophotometric
reflectance curve shapes. With practice one can become very skilled.
However, the exact shape of the reflectance curve depends on the concen-
tration at which a colorant is used. The advantage in plotting KIS on a log
scale for purposes of identification was pointed out in the previous section
in the discussion of pigment fading or darkening. When absorption curves
are plotted on a logarithmic scale, the shape becomes independent of the
colorant concentration. However, plotting absorption requires learning to
read curves "upside down" from what they appear to be in the commonly
measured reflectance form. This mental inversion may prove a bit difficult.
The problem can be alleviated by plotting l/(K/S) on a log scale-for
example, plotting 1/(K/S) versus wavelength on semilog paper-or by
plotting KIS on an "inverted" log scale-that is, with negative values
increasing up the ordinate and positive numbers increasing downward.
To illustrate what curves look like when plotted in the vari-
ous ways, curves for samples with five different concentrations of car-
bazole dioxazine violet in mixture with white (see fig. A.25) were plotted
as follows: diffuse reflectance (total reflectances of fig. A.25 minus 4%
first-surface reflectance); KIS; ll(K/S); log[l/(K/S)]; and log(K/S) against
wavelength. These are illustrated in figures 4.32-4.36, respectively.
Curves of the diffuse reflectance (fig. 4.32), of K/S (fig. 4.33), and of
l/(K/S) (fig. 4.34) all vary in resolution and shape with the concentration
of the colorant. Note that the maximum on the KIS curve (fig. 4.33) cor-
responds to the minimum on the reflectance curve but that the maximum
on the l/(K/S) curve (fig. 4.34) corresponds to the maximum on the
reflectance curve.
The use of a logarithmic scale, as in figure 4.35, illustrates
the parallelism between the curves in the key absorption region between
540 nm and about 640 nm. Because log[l/(K/S)] was plotted, the maxi-
mum corresponds to the maximum of the reflectance curve. On the right
side of figure 4.35 are listed values of the inverse function, log(K/S), with
negative numbers increasing up the ordinate from zero and positive val-
ues increasing downward. Figure 4.36 is similar except that values of
log(K/S) are on the left ordinate, and those of log[l/(K/S)] are on the
right. Log(K/S), with the negative logarithms increasing up the ordinate
from zero and positive log(K/S) values increasing downward, as illus-
trated in figure 4.36, is the scale that was plotted using the R-Cam on
the General Electric Recording Spectrophotometer (Derby 1952). This
was a very useful aid in identifying colorants and their approximate con-
centrations before the advent of computers.
The curve shapes illustrated in figures 4.35 and 4.36 in the
regions below 540 nm are not parallel because of the significant absorp-
tion of the rutile Ti02 in this region, where the carbazole violet has its
minimum absorption. Even on the reflectance curve, there is an indica-
tion of the carbazole violet absorption when the concentration is very
148
Figure 4.32
Reflectance curves of five concentrations
of carbazole dioxazine violet mixed
with Ti0
2
.
Chapter 4
100
90
80
70

60
:J
:t
:s
Cl)
u
c
l'l
50
u
Cl)
<;:::

.....
c
Cl)

40 Cl)
a..
30
20
10
400 450
y 0
carbazole dioxazine violet
500
mixtures in rutile Ti0
2
550
Wavelength (nm)
600
-----R----
650 700
low. However, the log(K/S)-or log[l/(K/S)] -curves show greater simi-
larities at most concentrations.
With computers, the transformations among functions of the
measured reflectance are relatively simple, and they are particularl y use-
ful when combined with curve-plotting capabilities. When curves are
generated from digital input, the resolution or visible detail depends not
only on the optical design of the instrument (primarily the resolution)
but also on the number of data points recorded and plotted. With some
modern instruments with excellent resolution, the number of data points
is so large-every 2 nm, for example-that a continuous curve is closely
approached. The advantage of a continuous curve, that is, of an analog
recording, is that it yields a maximum amount of information concerning
minute details of curve shape available from the instrument that might be
important for identification. For most colored materials measured in the
visible region, a 5 nm band-pass and plotting interval are the largest that
Figure 4.33
Curves of figure 4.32 plotted in terms of
K/S. Note that the KIS maximum corre-
sponds to the reflectance minimum.
Colorant Mixture 149
- V--11*C-- B -->--- G y 0 14-----R----
2.50
2.25
2.00
1.75
1.50
~ l v 1.25
1.00
0.75
0.50
0.25
0.00
400 450
carbazole dioxazine violet mixtures in rutile Ti02
500 550
Wavelength (nm)
600 650 700
would be safe to plot in order to yield reflectance curves that would be
unambiguous in their identification. (Transmittance measurements of
some transparent materials require higher resolution.)
In summary, transforming the reflectance to log(K/S) or
log[l/(K/S)] can be very helpful in identifying colorants based on mea-
surements and comparisons of their spectral reflectance curve shapes.
Reflection density
Many spectrophotometers designed primarily for measuring the transmit-
tance of nonscattering transparent materials record spectral curves in
terms of the optical density as well as in terms of percent transmittance.
Optical density, or absorbance, is defined as log( l/T) or log(/
0
/I), where
Tis transmittance, lo is the intensity of the incident beam, and I is the
intensity of the transmitted beam. In transparent samples, the log of this
150
Figure 4.34
Curves of figure 4.32 plotted in terms of
1 l(KIS). Note that the 1 l(KIS) maximum
corresponds to the reflectance maximum,
so that the shapes more closely resemble
the shapes of the reflectance curves.
Chapter 4
y 0 ----- R----
50
45
carbazole dioxazine violet mixtures in rutile Ti02
40
35
30
~ ~ 25
20
15
10
5
2.50
400 450 500 550 600 650
Wavelength (nm)
ratio is theoretically proportional to concentration and thickness:
log(I
0
/I} = CXA. As described in the first section of this chapter, C is
700
the concentration, Xis the thickness, and A is the unit absorption
coefficient. This relationship is described by the Beer-Bouguer equation.
As is pointed out in the next section, when S approaches zero, the
absorption coefficient A differs from the Kubeika-Munk absorption
coefficient K by a factor of 2-that is, A is half of K.
Many transmission spectrophotometers were also equipped
with an attachment for measuring reflectance. If reflectance is plotted
versus wavelength, this is fine. However, many workers recorded optical
density-in this case, log( 1/R) for reflected light-instead of reflectance
itself, mistakenly assuming that optical density was proportional to the
concentration of an absorbing material. In certain instances this is true,
but it is not generally so for materials with significant scattering.
Figure 4.35
Curves of figure 4.34 plotted as
log[1 !(K!S)]. On this basis the curves are
in general parallel in the characteristic
absorption region 530-650 nm. (The
short-wavelength differences, as well as
the differences at 660- 700 nm, are due to
the absorption of the rutile Ti0
2
in the
mixture.)
Colorant Mixture 151
v 8--..r--G y 0 ----R----
2.8
carbazole dioxazine violet mixtures in rutile Ti0
2
-2.8
2.4 -2.4
2.0 -2.0
1.6 -1.6
C)
....._
V)
;:;.
;;;:
::::.
0.0
1.2 -1.2
0.0
0 0
_J _J
0.8 -0.8
0.4 -0.4
0.0 0.0
-0.4 +0.4
450 500 550 600 650 700
Wavelength (nm)
An instrument developed by and widely used in the printing
industry for monitoring the color and intensity of the printing inks
applied to paper is the densitometer. This instrument measures the opti-
cal density of the reflected light-that is, log(l/R). In the case of layers of
transparent (nonscattering) inks applied over a reflecting substrate, this
reflection density is sometimes proportional to the product of the concen-
tration times the thickness, as was pointed out earlier in this chapter in
the section on glazes. Thus, the use of log(l/R) is valid when all of the
scattering comes from the substrate and the ink does not penetrate the
substrate-the paper, in the case of the normal printing process. Where
the ink penetrates the paper, two layers that both scatter are formed-
paper and ink, and paper alone-and the complex Kubeika-Munk equa-
tion (eq. 4.15), a hyperbolic equation, must be used.
For opaque materials that scatter as well as absorb, the use
of optical density is no longer valid-that is, the value of log(l/R) is
not proportional to the concentration of the absorbing substance. The
152
Figure 4.36
Curves of figure 4.33 plotted as log(K/ 5) ,
the "upside down " version of figure 4.35.
Chapter 4
v B G R
+0.8 --0.8
+0.4
2.5
--0.4
0.0 0.0
--0.4 +0.4
--0.8 +0.8
c;
~ ~
......
~
-1.2 +1 .2
Oil
Oil
0
0 _J
_J
-1.6 +1.6
-2.0 +2.0
-2.4 +2.4
carbazol e dioxazine violet mixtures in rutile Ti0
2
-2.8 +2.8
450 500 550 600 650 700
Wavelength (nm)
Kubeika-Munk relationship introduced earlier, (1 - R)
2
/2R = K/S,
should be used. The value of K/S is proportional to the concentration
of the absorbing material when scattering is constant, suitably weighted
for the individual absorptions and scatterings of each component in the
system. Throughout the paper industry, the Kubeika-Munk equations
are widely used. For example, Van den Akker (1949) in the United
States and Stenius (1951, 1953) in Sweden were two of the early pio-
neers in studying and developing the Kubeika-Munk equation appli-
cations in the paper industry, showing that two constants, one for
absorption and one for scattering, had to be used to describe the opti-
cal characteristics of paper. If the scattering is a constant for a given
type of paper, the scattering coefficient can be set equal to 1 at all
wavelengths, and single-constant theory based on KIS can be used as
described earlier in this chapter.
To show the relationship between K/S and log( 1/R), the
values of each were calculated for a range of reflectances. Figure 4.3 7
illustrates the relationship between reflectance and K/S, and between
reflectance and log(l/R). Since the curve of log(l/R) is much flatter than
that of K/S, the log(l/R) relationship, even if it were valid, would be a
far less sensitive measure in terms of curve shape than the one based on
KIS. This is illustrated in figure 4.38 for 2.5 % carbazole dioxazine violet .
Figure 4.37
Comparison of the reflectance as a func-
tion of K!S, which is (1 - R)
2
/2R, and the
reflectance as a function of log(1 / R) [cal-
culated as log(/
0
11)]. which is often taken
to represent the concentration of scatter-
ing materials.
Colorant Mixture
10.0 r---"""T""--.,..----.---...----.---.,..----.---.,..----,...----..,
' 1.0 '

'

b/:J
_Q
0 0.1
V)
'
.01
.001
0
'
'
'
20
'
'
'
--- K! S
'
1
- - - logR.
'
'
'
'
'
40 60
'
Percent reflectance
'
'
'
'
'
80
\
\
\
\
\
\
100
The lack of resolution in the absorption region 520-650 nm limits the
amount of information on curve shape in that region.
Talsky and Ristie-Solajie (1989) describe a technique for
using the fourth derivative of the log of l/R-d
4
(log R-
1
)/dl
4
-for dis-
tinguishing among the yellow pigments PY 1, PY 3, and PY 97 (see
fig. A.20) and the orange PO 1 in artists' paints. They all are Hansa
types ( acetoacetanilides).
153
154
Figure 4.38
Comparison of the data for the sample
with 2.5% carbazole dioxazine violet seen
in figure 4.32, plotted as log(K/5) and as
log[log(1 /R)J.
Chapter 4
+0.4
+0.3
+0.2
+0.1
&?: 0.0
---
b.O
0

-2 --0.1


b.O
.3 --0.2
--0.3
--0.4
I
I
I
I
I
0
I
log[log (1 / R)J
-----R----
\
\
\
\
'
'
'
'
'
...
2.5% carbazole dioxazine violet mixture in rutile Ti0
2
--0.5
-0.6
400 450 500 550 600 650 700
Wavelength (nm)
As has been stated, in most of the paper industry worldwide,
as well as in the paint, plastics, and textile industries, the Kubeika-Munk
relationship is used.
Special scales and methods used in industry
Many manufacturing industries-those representing textiles, paints,
plastics, and paper, to name a few-have developed special methods for
describing the color and appearance of their products and raw materials.
Some trade organizations representing these various industries have sci-
entific committees that establish their own standard methods to be used
in their industry. Examples of such trade organizations representing raw
material manufactures are the American Association of Textile Chemists
and Colorists (AATCC) (Celikiz and Kuehni 1983) and the Technical
Association of the Pulp and Paper Industries (TAPPI). Other industries
establish standard methods in conjunction with the American Society for
Colorant Mixture 155
Testing and Materials (ASTM), whose extensive methods encompass
numerous raw materials and finished products. The society's efforts are
overseen by specialized committees. For example, raw materials and
products for paint are represented by ASTM Committee DOl, for plas-
tics by ASTM Committee D20. Standards for artists' materials are a
part of Committee DOl and are under the jurisdiction of Subcommittee
DOl.57. Other manufacturing organizations utilizing many different
products also have their own methods. An example is the Society of
Automotive Engineers (SAE). Of the many industry methods, those of the
paper industry are of particular interest to conservators of art materials
on paper and of library materials.
Applications in the paper industry
Many of the studies made on paper involve the description of the bright-
ness of paper, which is an appearance term, and of the yellowness of
paper, either in terms of its yellowness as a visual attribute or as a mea-
sure of the concentration of the chromophoric groups causing the yellow-
ness. In the United States, TAPPI brightness is often employed as a
specification, as described in TAPPI Method T452 om-92 (1992c) and in
ASTM Standard D 985. In this method, brightness is equated with a lack
of yellowness; it is measured on a colorimeter in terms of reflectance as
measured through a special filter that has its maximum transmittance in
the blue spectral region at 457 nm.
It is logical to ask, "Why a special filter? Why not use the Z
tristimulus filter on a tristimulus colorimeter?" The answer is interesting.
Highly reflective whites, particularly those containing Ti0
2
pigments,
which absorb in the short-wavelength violet region of the spectrum, rep-
resent one case in which the CIE 1931 standard observer color-matching
functions do not correctly represent visual evaluations of whiteness
(Judd and Wyszecki 1975) . Jacobsen (1948) had earlier pointed this out.
The z color-matching function peaks at 445-446 nm. On near-white
samples that absorb in the violet, too much of the blue-violet reflectance
is eliminated by the Z filter of the colorimeter. By measuring with a filter
with its maximum transmittance at a slightly longer wavelength, 457
nm, more of the blue-violet light was included in the measurement,
which, the paper industry found, agreed better with visual evaluations.
Hence the development of the particular filter used to measure TAPPI
brightness came about. Judd and Wyszecki (1975) give an interesting
history of this problem, which ultimately helped to bring about the
establishment in 1964 of the 10 field standard observer functions by
the CIE. (The original 2 field of the 1931 color-matching functions for
the standard observer are equivalent to observing distant colors; those of
the 10 field of the 1964 standard observer represent viewing a material
at arm's length.) The 10 standard observer color-matching function for
the Z primary still peaks at 445 nm. It does give better estimates of yel-
lowness than the 2 standard observer, but it is not as good as the TAPPI
457 nm filter. Moreover, by 1964 the filter characteristics used in the
TAPPI brightness instrument were so well established that any change,
even if it were for the better, would be disruptive to the industry. Also,
156 Chapter 4
making a filter-detector-light-source combination that fits to a specific
curve shape for filter colorimeters is not an easy task. So TAPPI bright-
ness remains an industry standard. Although this standard is still widely
used, recent work has indicated that TAPPI brightness may not be the
best indicator of whiteness of paper (Jordan and O'Neill 1991). This is
discussed at greater length in the next section, on single-number color
scales for whiteness.
Today many workers with spectrophotometers use the
reflectance measured at 460 nm (or 457 nm) in place of the TAPPI filter
to produce values that correspond closely to TAPPI brightness. But note
that the definition of TAPPI brightness also requires a special illuminat-
ing and viewing geometry of 45/0, different from the sphere geometry
used in many spectrophotometers. In the author's limited experience,
the diffuse measurement using sphere geometry (specular excluded) and
monochromatic light with a spectrophotometer agreed moderately well
with the TAPPI brightness meter. On a difference basis, which is often all
that one is interested in, agreement was good.
Although the number is called "brightness," it is actually a
measure of yellowness. As a result, TAPPI brightness came to be used to
indicate the relative concentrations of chromophoric groups causing yel-
lowness; this is done by converting the reflectance to K!S, particularly to
indicate the change in yellowness due to exposure to various conditions.
In this form, it has been called the Post Color Number, or simply the P.C.
No. It is calculated from the TAPPI brightness expressed as a decimal
fraction and converted to K!S according to TAPPI Method T260 om-91
(1992a), as described in the following equation:
P.C. No.= 100 [cK I S)t -(KI S)to l (4.33 )
where subscript t denotes the reflectance measurement of the exposed
sample, and subscript t
0
denotes the reflectance measurement on the
unexposed original material. K!S is based on the TAPPI brightness or
on a spectral reflectance measurement at 457 nm (or 460 nm), both
expressed as a decimal fraction. The change in yellowness is sometimes
expressed as If positive, yellowing has occurred; if negative,
bleaching of the yellowness has occurred.
Note that the TAPPI brightness is a single-number color
scale. As will be discussed in the next section, single-number color scales
should be used only when the dominant wavelengths of the materials
being compared are similar and there is no metamerism between materi-
als. Note also that the P.C. No. application of the Kubeika-Munk rela-
tionship is based on single-constant, simple, colorant-mixture theory as
described earlier in this chapter. Therefore, the assumption is made that
the scattering coefficient, S, is constant. This is generally true when a
particular paper is being considered without the addition of any material
or treatment that alters the scattering of the paper. Many of the treat-
ments that may alter the scattering of the paper involve physically rein-
forcing the structure of old paper. Some methods of acid neutralization
involve the addition of neutralizing chemicals that result in the forma-
Colorant Mixture 157
tion of metallic oxides, such as ZnO or magnesium oxide, or of calcium
carbonate. These will scatter light and lighten the color of both the
paper and the absorbing materials in or on the paper. Single-constant
theory is, therefore, in this instance inappropriate to describe any
changes in the concentration of chromophoric groups, described by the
absorption coefficient, K. Thus, the single-constant, simple application
of the Kubeika-Munk coefficient, K!S, should not be used in such
situations. Berndt ( 19 89) discusses the limitation of single-constant
Kubeika-Munk calculations at greater length and provides an overview
of the use of the Kubeika-Munk equation in describing the optical char-
acteristics of pa per.
Other single-number color scales
Other widely used single-number color scales apply only to near-white
reflecting materials or to near-neutral transparent materials. On the sub-
ject of near-white reflectance description, two types of scales, in various
forms, may be encountered: (1) whiteness indices, and (2) yellowness
indices. Either may also be applied to paper products.
Perhaps a whiteness index varies most because of the ques-
tion of whether whiteness primarily favors blue whiteness or neutral
whiteness. In ASTM Standard E 313, a whiteness index was until
recently calculated as
WI=4B-3G (4.34)
where B and G are the tristimulus colorimeter filter measurements (B
for blue and G for green) or, based on CIE ( 19 31 ), the tristimulus values
(B = 0.847Z and G = Y). This index favors blue whites. In this same
method, a yellowness index was calculated as
YI= 100 (1 - BIG) (4.35)
The plastics industry introduced the following yellowness
index, which is different from equation 4.35 (ASTM Standard E 313 ).
This yellowness index was intended to be used for both reflecting
samples and for transparent materials.
YI= (100(1.28X - 1.06Z)]/Y (4.36)
where X, Y, and Z are tristimulus values for the 1931 standard observer
and illuminant C. Coefficients for other illuminants and the 10 standard
observer are given in ASTM Standard E 313.
In 1986 the CIE completed field tests on an equation for an
index of whiteness, W, and an accompanying equation for the tint, T w
(greenish or reddish), of near-white colors. The CIE equations, published
the same year, are
W = Y +800(x
0
-x)+1700(y
0
-y) (4.37)
Tw =1000(x
0
-x)-650(y
0
-y) (4.38)
where Y is the Y tristimulus value of the sample, x and y are the chro-
maticity coordinates of the sample, and x0 and y0 are the chromaticity
158 Chapter 4
coordinates of the perfect white diffuser, all for illuminant D65 and the
1931 2 standard observer (x
0
= 0.3127, y
0
= 0.3290). For the 1964
10 standard observer, the equations are unchanged except that the first
coefficient in the tint equation is reduced from 1000 to 900 (x
0
= 0.3138,
y
0
= 0.3310). ASTM Standard E 313 recommends the CIE equations
shown in equations 4.37 and 4.38.
Note that whiteness is measured as a difference from the
coordinates for the perfect white diffuser, for which W = 100 (the Y tri-
stimulus value). The higher the value of W, the greater the indicated
whiteness. The tint is equal to zero for a yellowish tint. The more posi-
tive the value of T W> the more greenish is the tint; the more negative, the
more reddish. The equations are to be used for W greater than 40, and
for T w between +3 and -3.
Since 1986 a number of papers on the CIE indices have
appeared, and they have been accepted as standard by many users. A
recent publication by Jordan and O'Neill (1991) concerning the visual
evaluation of the whiteness of paper indicates that the CIE whiteness
index, W, correlates best with the visual evaluation of whiteness. One of
the major aspects of their report is that their series of test papers used
for visual evaluation included many containing blue fluorescent whiten-
ing agents (FWAs).
For a concise summary of a number of whiteness indices
that have been proposed, Wyszecki and Stiles (1982) may be consulted.
For transparent liquids, many single-number color scales
exist primarily for materials that are near-white or yellowish in color.
A review of these many scales was given in Johnston's summary of the
work of the Inter-Society Color Council (ISCC) Subcommittee 14 on the
use of single-number transparent standards for yellowness (Johnston
1971a). Many liquid comparison standards have been used for products
as diverse as water, oils (vegetable and mineral), liquid chemicals, and
so on. Most of these are meant to be visual comparison standards, but
many of the scales have been converted into color-measurement terms,
as described in the report. Such single-number color scales are difficult
to use visually; the comparisons involve making matches that may be
severely metameric and that involve small differences in hue or dominant
wavelength. Nevertheless, they have served as a convenient, inexpensive,
and quick means for determining significant characterizations of relative
depths of color. That they are still in use is attested to by a recent publi-
cation on the use of the Gardner color scale for describing the yellowness
of resins and adhesives in the paper industry (Huebner and Moock
1992). (See also ASTM Standards D 156, D 1209, D 1500, and D 1544.)
The results obtained with such reference scales are subject
to a significant degree of uncertainty, but the scales do serve a useful
role in commerce by at least separating the "best" product color from
the "worst."
Chapter 5
Color in Specular (Mirror-Type) Reflection
Color of Metals (Nondielectrics)
In the section on transparent materials in chapter 4, equation 4.3
described the method for calculating the surface or interfacial reflec-
tance-that is, the specular reflection-for dielectric materials on the basis
of their refractive indices when the incident light is near normal (perpen-
dicular to the surface). The corresponding equation for nondielectric
materials (such as metals) is more complex as shown by the following:
P=[n(l-jk)-1]
2
n(l- jk)+ 1
at each wavelength
where the quantity n(l - jk) is the complex refractive index, j is the
imaginary constant -v=I, k is the absorption index, n is the real refractive
index, and 1 is the refractive index of a vacuum or, as generally used, of
air (Andrews 1960). Again this equation is valid only when the incident
light is near normal. When the absorption index, k, is nearly identical at
all visible wavelengths, the reflectance does not change significantly with
wavelength. But when k changes with wavelength in the visible region,
the reflectance also changes with wavelength, and the metal is seen as
colored. For metals such as copper, bronze, gold, and so on, k is not con-
stant at all wavelengths in the visible region, so the specular reflection
(gloss) is colored.
It is important to remember that the color of metals is due to
selective reflection at the air-metal interface. Thus, if a piece of highly
polished gold is measured with an instrument that measures only the dif-
fuse reflectance-that is, an instrument that excludes the gloss-the mea-
sured reflectance is near zero. Many color-measuring instruments are
designed to measure color only on dielectric materials and are inappro-
priate for use in measuring the color of metals. A detailed discussion of
the various types of color-measuring instruments is presented in chapter
9. In summary, instruments using 0/45 (illumination/viewing angle)
geometry, or the reverse, 45/0 geometry, are inappropriate for measur-
ing the color of metals, as are integrating sphere instruments that have
no provision for including the specular reflection. Ideally, for studying
the color of metals, an instrument that incorporates the option for
excluding or including the specular reflection-that is, for measuring
160 Chapter 5
the diffuse reflectance only or the total reflectance-is desirable. This
flexibility is particularly helpful in studying the smoothness of the surface
or the degree of polish.
If the metal to be measured is not perfectly polished-and few
metals are perfect mirrors-some of the light is reflected at the surface at
angles other than the specular angle. If the specular reflectance is colored,
regardless of the angles, however, this scattered surface reflectance is
probably also colored, though lower in chroma. A measure of the total
reflectance is required to describe the color. However, a measurement of
the nonspecular reflectance provides an indication of the perfection of the
surface polish. The ratio of the nonspecular (diffuse) reflectance to the
specular reflectance provides an index for the degree of smoothness or
polish. The specular reflectance, designated here by S, is calculated from
the two types of reflectance measurements: diffuse and total reflectances,
abbreviated SCE (specular component of the reflectance excluded) for the
diffuse reflectance, and SCI (specular component included) for the total
reflectance. (Some color-measurement workers jokingly refer to the two
measures as SEX and SIN, respectively.) Thus, the specular reflectance, S,
is determined by subtracting the SCE reflectance measurement from the
SCI measurement at each wavelength across the visible spectrum. For a
perfect mirror, the ratio of the diffuse reflectance, SCE, to the specular
reflectance, S, is zero at all wavelengths. The ratio SCE/S is an obvious
way to refer to the degree of diffuse reflectance of a metal surface, which
is an important aspect of the metal's appearance. The ratio is also a sensi-
tive indicator of change following exposure. One of the first changes that
occurs on metals when they are exposed to the elements is an increase in
the diffuse reflectance-that is, an increase in the ratio SCE/S. It has been
found that this ratio is a more sensitive measure of change than visual
evaluations alone.
The color of metal alloys is influenced by the amounts of the
various metals present. Some metals affect the hue of the alloy; others
affect only the chroma, or color vividness, by introducing a neutral,
dilution effect.
Measurements made on a series of bronze samples prepared
by W. Thomas Chase (1982) at the Freer Gallery of Art in Washington,
D.C., provide an intriguing start toward a study of a possible characteri-
zation of the composition of bronzes on the basis of their color. Samples
were prepared that contained varying amounts of copper, tin, and lead.
Before spectrophotometric measurements were made, the samples were
freshly polished. A Color-Eye Model 1500 was used for the reflectance
measurements at 20 nm intervals. The instrument is equipped with an
integrating sphere for measuring total reflectance (SCI) or diffuse reflec-
tance (SCE). The specular angle is about 8 from the normal. Both t ypes
of measurement were made on ten bronze samples and on the one copper
sample. Each measurement requires only a few seconds, which is an
important feature of the instrument because freshl y polished copper
begins to change color almost instantly if not protected immediately from
the air (by a lacquer, for example). Bronzes also change color quickly.
Table 5.1 presents a summary of the composition (Cu/Sn/Pb
proportions) for the ten bronze samples and one copper sample; the CIE
Sample Cu/Sn/Pb x
A 100/0/0 56.96
B 901713 69.34
c1
80/10/10 69.85
c2
old 65.33
D 77.5/22.5/0 63.15
E 76.5/12 .6/10.9 63.19
F 75/15/10 63.35
G 70/27.5/2.5 61.21
H 65/2 .5/32 .5 66.44
(not provided) 63.18
90/10/0 65.33
Color in Specular (Mirror-Type) Reflection 161
tristimulus values and chromaticity coordinates calculated for illuminant
C and the 19 31 standard observer from the total reflectance measure-
ments; the Munsell notation calculated from the CIE coordinates; and the
ratio of the Y tristimulus value measured with the specular reflectance
excl uded (SCE) to the Y value for the specular reflectance (S) alone.
The spectral curves of the total reflectances are presented in
figures 5 .1 and 5 .2. Figure 5 .3 shows the CIE chromaticity diagram for
the total reflectance data from table 5.1; lines of constant dominant
wavelength (corresponding approximately to constant Munsell Hue) and
Total reflectance Munsell notation
y
z x y Hue Value/Chroma
YscEfys
51.57 36.85 0.3918 0.3547 3.01YR 7.55/5.44 0.277
67.73 54.31 0.3623 0.3539 7.73YR 8.47 /3.56 0.052
69.23 57.18 0.3559 0.3527 8.49YR 8.54/3.15 0.064
64.49 51.89 0.3595 0.3549 8.47YR 8.30/3.35 0.125
63.58 55.86 0.3458 0.3482 0.19Y 8.25/2.38 0.028
63.63 56.48 0.3447 0.3471 0.09Y 8.25/2.31 0.094
62.89 51.22 0.3570 0.3544 8.79YR 8.21/3.17 0.077
62.19 62.86 0.3286 0.3339 0.91Y 8.17/1.22 0.038
63.52 53.44 0.3623 0.3463 4.59YR 8.24/3.64 0.126
63.99 63.12 0.3320 0.3363 0.28Y 8.27 /1.45 0.102
63.88 50.58 0.3633 0.3552 7.82YR 8.20/3.63 0.020
' The ratio of the Y tristimulus value measured with the specular component of the reflectance excluded (SCE. diffuse reflectance) relative to the Y value for the specular component
of the reflectance, S (S = SCl -SCE) . The ratio indicates how near to a perfect mirror the sample appears to be; if perfectly mirrorlike, the ratio would be zero.
Table 5.1
Composition and colorimetric properties
of various bronze alloys (alloys provided
by T. Chase).
Figure 5.1
Total reflectance curves of bronze samples
(data for which are presented in table 5.1)
of various compositions as indicated.
Note that the curve for pure copper is
very different from those for the bronzes.
(The composition of sample I was not
provided.)
90
80
70
50
40
30
20
10
400
bronze samples (T. Chase)
Cu I Sn I Pb
...... A 100 / 0 /0
-0- D 77.5 I 22.5 I 0

H 65 I 2.5 I 32.5
-l::r
-D- 90 I 10 I 0
440 480 520 560 600 640 680
Wavelength (nm)
162
Figure 5.2
Total reflectance curves of bronze samples
of various compositions (see fig. 5.1 and
table 5.1 ), as indicated.
Chapter 5
lines of constant excitation purity (corresponding approximately to con-
stant Munsell Chroma) are included for reference.
A brief study of the data indicates that the Munsell Hue is
largely dependent on the tin/copper ratio (Sn/Cu) and that the addition
of lead (Pb) primarily lowers the Chroma. Figure 5.4 shows a compari-
son of the calculated Munsell Hue to the tin/copper ratio. No such
simple generalization can be made concerning the relationship of the
observed Chroma and the lead content, however. The Munsell Value
seems to be unaffected by composition for all the bronzes; only pure
copper is significantly different, being darker than the bronzes.
As mentioned, the ratio of the diffuse reflectance measure-
ment (SCE) to the specular reflectance measurement (S) is an indicator of
the relative amount of smoothness or polish-that is, of surface scatter-
ing. From the calculated data presented in the last column of table 5.1, it
can be seen that the copper had the highest ratio, indicating that it was
least mirrorlike. Indeed, when copper is polished, it immediately begins
to change, so that to keep its polish, copper needs to be protected from
the atmosphere.
In a program of study concerning the efficacy of clear coat-
ings in protecting copper from loss in gloss and from darkening, it was
found that an index of change based on the specular reflectance change
~
g
<I>
u
c
_:g
u
<I>
q:::
~
......
c
<I>
u
v
CL
100
90
80
70
60
50
40
30
20
10
400
bronze samples (T. Chase)
Cu I Sn I Pb
...... 8 9017 I 3
-0- c, 80/10/10
~ E 76.5 I 12.6 I 10.9
-6:- F 75 I 15 I 10
-D- G 70 I 27.5 I 2.5
440 480 520 560 600 640 680
Wavelength (nm)
Figure 5.3
Portion of the CIE 1931 chromaticity dia-
gram showing the location of the copper,
A, and bronze samples whose curves are
shown in figures 5.1 and 5.2. (data taken
from table 5.1 ). The open circle is the illu-
minant point for illuminant C. Lines radiat-
ing out from the illuminant point are lines
of constant dominant wavelength for 574,
580, 582, 584, and 588 nm. Lines perpen-
dicular to the dominant wavelength lines
are lines of constant 10%, 20%, and 30%
purity.
Figure 5.4
Munsell Hue plotted against the tin/ copper
ratio. With a few exceptions, there appears
to be a reasonable correlation between
the two.
Color in Specular (Mirror-Type) Reflection 163
cu
::J
I

c
::J
::?:
0.350
0.340
0.330
0.320
0.310
OY
10YR
9YR
8YR
?YR
6YR
5YR
4YR
Bronze samples (T. Chase)
total reflectance measured relative to BaS04
0.310 0.320 0.330 0.340 0.350 0.360 0.370 0.380 0.390
A
0.1

E

F
x
Munsell Hue vs. Sn/Cu ratio
(based on total measured reflectance)
Cu I Sn I Pb
A 100 I 0 I 0
B 90 I 7 I 3
C1 80 I 10 I 10
D 77.5 I 22.5 I 0
E 76.5 I 12.6 I 10.9
F 75 I 15 I 10
G 70 I 27.5 I 2.5
H 65 I 2.5 I 32.5
90 I 10 I 0
0.2 0.3 0.4
Sn/Cu
0.5
164 Chapter 5
and on the diffuse reflectance change following exposure provided
the most sensitive early prediction of failure of the coating to protect
the copper (Spindel 1973).
As with bronze, the color of other metal alloys also changes
with composition. The hue of gold, for instance, depends on the relative
amounts of other metals, such as copper or silver, in its composition. Even
alloys generally considered to be colorless, such as stainless steel and alu-
minum, may exhibit some color. For a review of methods for appearance
evaluation of metals, see ASTM Special Technical Publication 4 78.
Bronzing
The visual appearance of dielectric (organic) materials can occasionally
be metal-like, particularly following exposure. Such materials may
exhibit distinct color in the specular reflection, just as metals do. The
phenomenon is generally termed "bronzing." Bue and coworkers (194 7)
characterized two types of bronzing: one being metallic reflection from
the air-object interface, which they term "interface bronze," and the
other arising from light entering the surface and being reflected specu-
larly from closely adjacent layers, ridges, or particle layers just under the
surface; they term this type of specular reflection "interference bronze"
(see the next section for a further discussion of interference color). The
interference reflection causes the appearance generally called iridescence.
Either type of bronzing-though the two have different underlying
causes-owes its color appearance to the light reflected at the specular
angle. Distinguishing the difference between the two types, whether by
visual evaluation or by instrument measurement, involves changing the
angles of illumination and viewing: "interface" bronze (as from metals)
is the same hue regardless of the incident angle; "interference" bronze
(iridescence) changes in hue with a change in the angle of incidence. For
complete characterization of interference color, an instrument that mea-
sures spectral character using various angles of illuminant incidence and
viewing is desirable. Such an instrument is called a goniospectropho-
tometer, which has only recently become readily available among com-
mercial instruments.
Whatever the origin of the color in the specular reflection, the
use of an instrument that measures the color of the specular reflectance is
necessary. In a paper by Johnston-Feller and Osmer (1979), two types of
instruments were used to characterize the color of bronzed paint films:
a goniospectrophotometer and an integrating sphere spectrophotometer
capable of measuring both total and diffuse reflectance, from which the
color of the specular component could be calculated.
As a demonstration, masstone paint films were pigmented with
an azo red, PR 144, at 40% pigment (weight percent in the dry film); a
blue, PB 15:3, at 50%; a halogenated copper phthalo-
cyanine green, PG 7, at 45%; and a carbon black, PBk 7, at 38%. At these
concentrations all appeared "bronzy." Visual evaluations of the bronzed
films resulted in the descriptions given in table 5 .2. (The cover of the
Figure 5.5
Goniospectrophotometric curves of
blue masstone paint
panels. The upper two curves show the
bronzed color (a reddish purple) observed
at the specular angles of 15 and 30. The
lower, flat curves were obtained from a
panel coated with two layers of clear
vehicle, which eliminated the bronzed
color on the panel.
Color in Specular (Mirror-Type) Reflection 165
March 1979 issue of the Journal of Coatings Technology shows Robert
Feller's photographs of these bronzed samples.) Covering the bronzed films
with a clear, transparent coating decreased the bronzed appearance and,
when two such layers were applied, eliminated it completely.
The four samples were measured with the Trilac goniospec-
trophotometer at two specular angles, 15 and 30 (the geometries are
described as -15/15 and -30/30). One of the four, the sample made
with blue is illustrated in figure 5 .5. The specular
reflectance measured on the bronzed sample is purple-reflecting red,
blue, and violet wavelengths. The specular reflectance of the same
sample, when covered with two coats of clear vehicle, shows primarily
the surface reflectance of the colorless vehicle's gloss, reflecting the light
source at all wavelengths.
The specular reflectance was also determined from measure-
ments made using the same Trilac spectrophotometer operated in the
-v B G
y
0 R
100
90
80
70
60
50
40
30
20
10
400
sample: blue masstone
comparison: Ba50
4
no overcoat
- - - - - - 2 overcoats
--===== -- -15/15
.,,,.- --::::----
----:::::::--------------
-----
450 500 550
Wavelength (nm)
-----
-30130
600 650 700
166
Figure 5.6
Specular reflectance plotted as the differ-
ence between the specular reflectances of
the bronzed sample of phthalo blue mea-
sured with sphere geometry and measure-
ments of the same panel when overcoated
with a clear vehicle. The bronze color is
reddish purple.
Chapter 5
sphere mode. The differences in the specular reflectances of the bronzed
samples (subscript B) and the coated panels (subscript C) were calcu-
lated: (SCI - SCE)
8
- (SCI - SCE)c. The differences determined at each
wavelength are illustrated in figure 5.6. Although the shape of the curve
defined by the dots in figure 5 .6 and the curve shapes in figure 5 .5 are
not identical, they are similar in shape, both showing red and blue wave-
lengths in the specular reflection of the bronzed sample. It can be seen
from figure 5 .6 that in the green region, complementary to the purple
bronzing reflection, the difference of the bronzed sample and the coated
sample is negative-that is, the specular reflectance is less on the bronzed
sample than on the coated sample. This is frequently encountered when
bronzed samples are measured with sphere geometry and the specular
reflectance is calculated by difference.
The CIE chromaticity coordinates for the four bronzed
samples are plotted in figure 5.7. The open symbols refer to the measure-
ments made with the goniospectrophotometer at the two specular angles
illustrated in figure 5.5; the solid black symbols are calculated by the dif-
ference method by use of sphere geometry.
Agreement is apparent between the visual evaluation of the
bronzed samples (table 5.2) and the estimates of hue (described in fig. 5.7
by the dominant wavelength) and chroma (described in fig. 5.7 by the
distance from the illuminant point ). The bronze color of phthalo blue is
purple and strong; that of azo red is greenish yellow and strong; that of
+1
+ I-

1 1
-.



r ~

- I-
-1
400
I
500

Wavelength (nm)
I
600
-
700
Figure 5.7
Portion of the CIE 1931 chromaticity dia-
gram for illuminant C, showing the loci of
the samples measured by the goniospec-
trophotometric method at two specular
angles, 15 and 30 (open symbols). These
are connected by lines toward the illumi-
nant point for the samples overcoated with
clear vehicle. Solid black symbols are the
loci of the points calculated by the differ-
ence method from the measurements
made with sphere geometry.
Table 5.2
Visual description of bronzed paint
samples.
Color in Specular (Mirror-Type) Reflection
0.40 ~ . .
0
azo red
0.38
v
phthalo green
!::..
phthalo blue
D
carbon black
-- -30/30
0.36
- - - -15/15
0.34
"' 0.32
700
0.30
il luminant C
0.28
0.26
black symbols: sphere geometry
0.24 ..._ ___ _._ ___ __.._ ___ __., ____ .__ ___ _..._ ___ _.
0.28 0.30
Masstone sample
Phthalocyanine green
Azo red
Phthalocyanine blue
Carbon black
0.32 0.34 0.36 0.38
x
Visual description of bronzing
blue, weak
yellow, strong
purple-red, very strong
yellow, weak
0.40
167
phthalo green is blue-violet and weak; and that of the carbon black is yel-
low and weak. These properties are consistent between the two types of
measurements: goniospectrophotometric measurements and the "sphere
difference" calculation of the specular reflectance.
The points plotted near the illuminant point are the variable
angle measurements made on the samples with two coats of clear vehicle
added. (If the reader is disturbed by the location of these points away
from the illuminant point, the explanation is as follows: the variable
angle values were measured with a 20% neutral-density filter, with the
result that any slight errors have been multiplied by 5; in contrast, in
the case of the sphere measurements, the amount of specular light deter-
mined was no ta bl y small, generally less than 10 % . )
168 Chapter 5
It is remarkable that the chromaticity agreement shown in
figure 5.7 is so good in both dominant wavelength (hue) and purity
(chroma), despite the extreme differences in the magnitude of the mea-
surements. The results serve as an illustration of the sensitivity that can
be achieved in measurements of the specular component by taking the
difference between the total reflectance and the diffuse reflectance using
an integrating-sphere instrument. That the bronze colors measured with
-15/15 and -30/30 geometries were similar does not necessarily mean
that the type of bronzing was not due to interference effects. In fact, the
differences may be significant. Bue and coworkers ( 194 7) point out that
interference bronzing changes most radically at lower angles of incidence
(60 or greater). They add that the color of the observed bronzing may
also change following continued exposure.
The reader may well wonder why this phenomenon is dis-
cussed in these pages dealing with the color of museum objects and mate-
rials. How many times has bronzing occurred and been described? It is
difficult to know the answer. However, there are two primary reasons for
pointing out the occurrence of this phenomenon: ( 1) Unless the observer
is attuned to looking for bronzing, it may go undetected, particularly
when the effect is very small. (2) Many of the newer pigments introduced
since 1930 have a greater tendency toward exhibiting this behavior than
do most classical pigments. Hence, bronzing may occur more frequently
in modern paintings and particularly in modern inks (Smith 1963 ). One
of the traditional pigments subject to bronzing is iron blue (ferric ferro-
cyanide, or Prussian blue), used by artists for over a century. It exhibits
interference bronze; with angular variation, it changes in color from red
to green to blue.
Perhaps the most widely used modern pigment subject to a
subtle specular reflectance change, particularly following exposure, is
rutile Ti0
2
. Until it is pointed out, the selective color of the specular
reflection is often not noted visually. Even though the surface gloss may
appear to be unchanged (that is, the reflection is not more diffuse), the
spectral distribution of the specular reflection may run the gamut of
spectral colors (violet, blue, green, yellow, red, purple)-albeit of low
chroma-subtly affecting the visual color. Armstrong and Ross (1966)
attribute this phenomenon to interference effects resulting from the
microstructure in the rutile paint films. Whether this phenomenon has
been observed in modern paintings or on painted objects or buildings is
not known, but the reader is forewarned to make careful visual examina-
tions of the color of the gloss (the specular reflection).
In the article by Johnston-Feller and Osmer (1979), a num-
ber of pertinent references to the subject of bronzing are given. The
possibility of the occurrence of this phenomenon is a further reason
for measuring colors with the specular component of the reflectance
included (total reflectance, SCI) as well as excluded (diffuse reflectance,
SCE). The character of the chromaticity of the specular reflection alone
can then be analyzed.
Color in Specular (Mirror-Type) Reflection 169
Pearlescence and Iridescence
Both pearlescence and iridescence result from the structure of materials,
which is often flakes or thin layers with a refractive index higher than
that of the surrounding medium. The distinction between pearlescence
and iridescence is the nature of the specular light reflected. If at any spec-
ular angle the observed reflectance is white, the effect is called pearles-
cence or luster; if the specular reflectance is colored (chromatic) and
changes hue when the illuminating and viewing angles are changed, it
is called iridescence or interference color. In either case, the specular
reflection occurs within the object or film and is due to the structure of
the object and not to absorbing colorants-dyes or pigments-incorpo-
rated into the object.
Numerous examples of interference colors occur in nature:
the iridescence of feathers on certain birds, butterfly wings, certain
beetles, and some tropical fish; the iridescence arises from structural color
and not from pigmentation. The myriad colors of soap bubbles and oil
slicks are also familiar examples. As described in the previous section on
bronzing of pigmented dielectric films, one type of bronzing (metallic
appearance) is attributable to interference and is characterized by changes
in color as the angles of illumination and viewing change.
The interference colors so characteristic of many Tiffany glass
objects are also illustrative of the phenomenon. Many ancient glasses,
buried in the earth for centuries, exhibit beautiful interference colors in
the surface that developed during their long burial.
Interference color results when incident light is reflected from
a regularly organized structure within an object that can reflect light of
certain wavelengths in a particular way: When the reflected light is in
phase with one particular wavelength of the incident light, it reinforces
that wavelength and eliminates other wavelengths that are out of phase.
Figure 5.8 illustrates this phenomenon. Each line in the figure represents
waves of the same wavelength, and each pair illustrates different ways in
which two waves may interact. Wavelength is the distance from one crest
to the next or from one trough to the next. In the upper pair, one wave
has its crest where the other has its trough-the two waves are com-
pletely out of phase. The result is complete cancellation of light, that is,
darkness. In the middle pair, the two waves partially cancel-that is, they
are partially out of phase, with the result that the reflected intensity is
very much weakened. In the lower pair, the waves are completely in
phase and reinforce one another; the result is a pure, strong color associ-
ated with that particular wavelength. The resultant color is pure spectral
color of a single wavelength.
The flash of specular color observed at certain angles in many
objects exhibiting interference color is most frequently due to laminar
structures, that is, layers of different refractive indices of such thickness
that interference of light can occur. Examples are mother-of-pearl, but-
terfly wings, and hummingbird feathers. Modern interference pigments
170
Figure 5.8
Comparison of the behavior of two light
rays of the same wavelength (distance
from crest to next crest, or from trough
to next trough). The upper pair are com-
pletely out of phase, so that one wave
cancels the other, resulting in darkness.
The middle pair are partially out of phase,
resulting in weakened light. The bottom
pair are completely in phase, and the
resulting color is pure and strong.
Figure 5.9
Light ray 2 enters the film at point A; is
bent toward the perpendicular (dotted
lines) when it enters the denser medium;
is specularly reflected at the second inter-
face; and, upon exiting, is bent away from
the perpendicular (ray 2'). Light ray 1 is
specularly reflected as ray 1' at the top
surface and is in phase with ray 2', rein-
forcing the color (n
0
is the lower refrac-
tive index; n
1
is the higher refractive
index of the denser medium; and X is
the thickness) (after Simon 1971 ).
Chapter 5
are likewise laminar structures (platelets or thin flakes). The iridescence
depends on two factors: the difference in refractive index between the
layers, and the thickness of the layers. Transparent materials in layers
reflect at the two interfaces where the refractive index changes: the top
and bottom of the layer or film, as described in the section on transpar-
ent materials in chapter 4. The amount of light reflected at the interfaces
depends on the refractive index difference and on the angle of incidence.
Figure 5 .9 illustrates how two beams that are incident on a film, one
reflected at the top surface of the film and the second from the bottom
surface, reinforce each other if the wave trains of the two beams are in
phase. In figure 5 .10, the optical thickness has been changed, and the
reinforced color has a different wavelength from the specularly reflected
light of the thicker film illustrated in figure 5 .9. The optical thickness
can also be changed by increasing the angles of illumination and view, as
ill us tr a red in figure 5 .11. Here the thickness is the same as in figure 5. 9,
but the angle of incidence and view is higher, so that the optical path is
increased, and the reinforced spectral color is different. (These illustra-
2 1' + 2'
B
Figure 5.10
The same phenomenon is occurring here
as in figure 5.9, but because the thickness
of the film is different, the wavelength of
light that is reinforced is different (after
Simon 1971 ).
Figure 5.11
The optical path length or thickness is dif-
ferent, because the angle at which the rays
strike the surface is greater than in figures
5.9 and 5.10, and the color is again differ-
ent (after Simon 1971 ).
Color in Specular (Mirror-Type) Reflection
2 1' + 2'
X n
1
-}
B
tions are adapted from Simon [1971].) Thus, the colors change con-
stantly as the illuminating and viewing angles are changed, producing
a beautiful display of very pure spectral colors in the reflection.
171
Other structures may also be responsible for selective specu-
lar reflection, as Simon (1971) points out and illustrates. An appropriate
arrangement of ridges or slits, such as in diffraction gratings, or an
appropriate spacing of minute particles in equidistant layers (a space
lattice) will also produce interference of light.
One may ask what happens to the wavelengths that are not
the same as the one that is specularly reflected? Because the layers that
produce the interference are transparent and nonabsorbing, the wave-
lengths of light other than the one specularly reflected are transmitted.
The result is that interference platelets have two characteristic colors, the
one that is specularly reflected and the other that is transmitted. The
transmitted radiant flux is always lower in purity than the specularly
reflected color. In most situations some of the light of both types may be
scattered-from the edges of platelets, for example, or from other materi-
als present. To retain the maximum specular reflection, the scattering
must be kept as low as possible.
B
172 Chapter 5
History of pearlescent and iridescent pigments
In the first of his chapters in the Pigment Handbook, Greenstein (1973a)
traces the commercial development of interference (iridescent) pigments
and pearlescent (also called lustrous or nacreous ) pigments for use in col-
oring man-made articles. He attributes the beginnings of this specialized
pigment industry to the introduction of plastics, notably the introduction
of nitrocellulose, developed about the time of World War I.
All of the pearlescent and interference pigments developed for
incorporation into synthetic polymers owe their specular reflectance to
thin platelets or flakes.
According to Greenstein, the first pearlescent pigments were
made from fish scales, primarily those of herring. The reflective platelets
consist of guanine (2-amino-6-hydroxypurine), with lesser amounts of
hypoxanthine (6-hydroxypurine). Because the amount of nacreous pig-
ment obtained is less than 0.1 % of the weight of the fish, it took a lot
of fish to make a pound of pigment. The price reflected this: in 1972 a
pound cost $90.91 to $145.45, depending on quality. Fish-scale nacreous
material has been used widely for many years in simulated pearls and in
cosmetics such as nail enamels. The platelets are sold as suspensions,
usually at 11 % or 22% concentration.
From the beginning of interest in pearlescent pigments for use
in plastics and coatings, attempts were made to create lustrous pigments
synthetically. The 1920s saw the development of mercurous chloride.
However, the use of this poisonous compound was short lived. In the
1930s there was the development of lead hydrogen phosphate (PbHP04 ),
but the luster was not very high compared with that of the natural
pearlescent pigments. PbHP04 had its maximum usage in the 1940s
and early 1950s.
During the 1940s and 1950s nacreous lead carbonate was
developed; it proved to be as lustrous as the natural pearlescents. By
varying the thickness of the crystals, interference pigments could be cre-
ated. Although nacreous lead carbonate was significant in the develop-
ment of synthetic pearlescent and interference pigments, its place has
now largely been taken by newer nonlead pigments with similar optical
properties. Although no known case of lead poisoning from pearlescent
buttons or pearls has been reported-the pigments are embedded in very
stable polymer matrices-the possibility that their use would be frowned
upon because of their lead content, even if uningestible, caused most
manufacturers to discontinue their production (Greenstein, personal
communication, 1993) .
Other lustrous, pearlescent pigments developed from the
1940s to the 1960s were lead hydrogen arsenate (not very important
commercially because of its toxicity) and bismuth oxychloride (BiOCl).
The latter is nontoxic. However, since BiOCl darkens on exposure to
sunlight, its use in applications other than cosmetics requires protection
by ultraviolet absorbers. It is lustrous but not appropriate for making
interference pigments.
Probably the most significant event was the development of
Ti0
2
-coated mica flakes in the 1960s. Controlling the size and thickness
Color in Specular (Mirror-Type) Reflection 173
distribution of the mica platelets and the thickness of the high-refractive-
index Ti0
2
coating makes it possible to create interference pigments cover-
ing a broad spectrum, as well as pearlescent flakes. Because these pigments
and flakes are nontoxic and durable, their use has been expanding.
The optical thickness of the coating on the surface of the
interference-pigment platelets determines their color. For highest luster,
the platelets range in size from about 10 to 40 m in their longest dimen-
sion; their thickness is considerably less, 0.06 to 0.8 m, depending on
the type. The critical dimension is the thickness of the interference layer.
For example, an optical thickness of less than 140 nm results in white
luster or pearlescence; above 200 nm optical thickness, the interference
light is colored. With increasing thicknesses, it varies from yellow to red-
gold to magenta to purple to blue to green at 370 nm optical thickness
(Greenstein 1988). Further increases in the optical thickness result in a
repetition of this order. Increasing the optical thickness in increments of
50% results in waves in phase (because the light travels in two direc-
tions, entering and exiting in the layer).
Figure 5.12, reproduced from Greenstein's second article
(1973b) in the Pigment Handbook, illustrates the change in the specular
reflectance with angle of illumination and viewing for a basic lead car-
bonate, green interference pigment as measured on the Trilac goniospec-
trophotometer. This instrument is described in chapter 1. As noted there,
the incident angle is designated as negative and the viewing angle as posi-
tive, because they are on opposite sides of the perpendicular. It can be
seen from figure 5.12 that increasing the angles from 15 to 45 results
in a change in the hue from green to blue. The explanation and equa-
tions that predict this change are presented by Greenstein (1973b).
As described above, one of the characteristics of all pearles-
cent and interference pigments is their transparency-that is, their lack of
absorption. In the case of interference pigments, the transmitted light is
complementary in color to that reflected. Figures 5.13 and 5.14, redrawn
from Greenstein's article ( 1973 b ), illustrate, respectively, the reflectance
and transmittance of a blue and a green basic lead carbonate interference
pigment. If one examines such pigments under the usual optical micro-
scope, the transmitted color is observed. This is the complementary color
of the specularly reflected color, as shown in figures 5 .13 and 5 .14. Note
in these illustrations that the transmitted color is much lower in chroma
than the specularly reflected color. If the microscope is arranged so that
there is vertical illumination in a dark field, the reflected interference
color can be observed.
In order for the reflected interference color to be seen alone,
the transmitted light must be absorbed. In polymer films, this can be
achieved by applying the film over an absorbing substrate, such as a black
or dark brown. Mother Nature utilizes this phenomenon in many cases of
iridescent color. For example, the underside of the brilliant blue Morpho
butterfly wing, a beautiful example of interference color, shows dark
brown pigmentation. The hummingbird exhibits its flashes of iridescence
only at the specular angle. At all other angles the little bird's feathers are
a dull brown, well camouflaged among the branches of a tree, because of
174
Figure 5.12
Shown here is the change in color in the
specular reflectance of a basic lead car-
bonate green interference pigment as
the angle of incidence is changed. At
the higher angle of incidence, -45,
the color is a blue instead of the blue-
green observed at -15 incidence (from
Greenstein 1973b. Copyright 1973
John Wiley & Sons, Inc. Reprinted by
permission).
Chapter 5
- v--:----- B ------- G
y 0
9
8
7
6
<!)
u
c

u
<!)
;:;::::
5
<!)
>

Qj
C<'.
4
3
2
400 450 500 550
Wavelength (nm)
600
-----R----
650 700
their absorption color. Thus, at angles other than the specular angle, one
observes the substrate color; the transmitted color has been absorbed.
If a thin film of a polymer containing an interference pigment
is applied over a highly reflective substrate, such as a white, the interfer-
ence color is still observed at the specular angle, but the chroma (purity)
of the observed color is decreased because of the inclusion of some of the
reflection from the substrate-diffuse light that is the color of the trans-
mitted light. Thus, over white, at angles other than the specular angle, the
color observed is the transmitted color diffusely reflected by the substrate.
The pigment industry soon realized that the perceived color
of pearlescent or interference pigments can be modified by the inclusion
into the polymer film of a transparent (low-scattering) absorbing pig-
ment, such as a black or a transparent chromatic colorant. Care must be
taken in selecting the transparent chromatic colorant to use in combina-
tion with an interference pigment. If a transparent pigment that absorbs
Figure 5.13
Specular reflection of a blue-reflecting
basic lead carbonate interference pigment
at an incident angle of -15. The second
curve illustrates the transmission color, a
low-chroma yellow-orange, the com-
plement of the blue reflection (from
Greenstein 1973b. Copyright 1973
John Wiley & Sons, Inc. Reprinted by
permission).
Color in Specular (Mirror-Type) Reflection
(lJ
u
c
.::l
u
(lJ
;;:::
1:!
(lJ
>

v
C<'.
90
80
70
60
50
40
30
20
10
400 450 500 550
Wavelength (nm)
0
600
175
i.---R---
100
90
80
70
60
v
u
c
<IJ

.E
"'
50
c

.....
.....
c
v

v
40
0...
30
20
10
650 700
all colors except red is used in combination with a red-reflecting inter-
ference pigment, the film of the mixture will appear red at all angles of
illumination and viewing, although the red specular reflection will be
enhanced by the inclusion of the absorbing pigment. Today some inter-
ference pigments of the Ti0
2
-coated-mica type are manufactured with a
layer of transparent absorbing pigment as well as the Ti0
2
layer; others
may be made by incorporating an absorbing pigment into the Ti0
2
layer.
When the absorbing pigment has the same hue as the interference color,
the highlight color observed at the specular angle is enhanced, and the
diffuse color seen at other angles is the same hue if the layer is applied
over a reflecting neutral substrate. In contrast, when the absorbing color
in the Ti0
2
layer is different from the interference color, a two-tone
effect can be created, as long as the two colors do not cancel each other.
An example of such a modified interference pigment would be a blue
interference pigment with a transparent red layer over the Ti0
2
layer.
176
Figure 5.14
Specular reflection of a green-reflecting
basic lead carbonate interference pigment
at -15incidence. The transmission color
is the complementary hue, a magenta of
very low chroma (from Greenstein 1973b.
Copyright 1973 John Wiley & Sons, Inc.
Reprinted by permission).
Chapter 5
90
80
70
60
Q)
u
c
~
u
Q)
q::
~ 50
Q)
>
:;:::;
<i:i
a:;
er:
40
30
20
10
400 450 500 550
Wavelength (nm)
600 650
90
80
70
60
Q)
u
c
<i:i
:t:
~
50
c
~
c
Q)
~
Q)
40
CL
30
20
10
700
Figure 5.15 illustrates goniospectrophotometric measurements of such a
modified interference pigment applied over a white substrate (Greenstein
1988). The transparent red in the anatase Ti0
2
layer is carmine, evident
by the two absorption bands at 520 and 560 nm (see fig. 4.19).
Because there are so few goniospectrophotometers available,
it is worthwhile to illustrate the characteristics of the same type of
pigment as measured by use of sphere geometry with the specular
reflectance included. Figure 5 .16 illustrates the spectral curves of the
same type of pigment measured over a white substrate and over a black
substrate. (The same type of spectrophotometer used in fig. 5 .15 was
used in the sphere mode to obtain the curves in fig. 5.16. ) The curve
measured over black, which absorbs the transmission of the blue inter-
ference pigment and that of the red absorption pigment, shows only the
blue interference color. The curve of the pigment applied over white
shows the combination of the blue interference color with the red
Figure 5.15
A blue interference pigment modified with
a red (carmine) transparent pigment layer
over the anatase Ti0
2
layer; the lacquer
film is applied over white. At incident and
viewing angles of -15/15, the blue spec-
ular reflectance color is observed. At a
viewing angle of 60 with the same -15
incident angle, the weak red transmission
color is combined with the reflection of
the carmine red pigment over white:
the color observed is a bright red (after
Greenstein 1988).
Color in Specular (Mirror-Type) Reflection
"'
u
c

t
"'
q::

"'
>

OJ
0::
8
6
4
2
R/B anatase-mica combination interference pigment
(gonio measurement over white)
0.8
0.6
0.4
0.2
0 0.0
400 500 600 700
Wavelength (nm)
absorption color; the result is a purple. The dominant wavelength of
177
the measurement over black is 4 73 nm, a blue; over white, it is 500c nm,
a purple. These curves and measurements are included to illustrate that
considerable information concerning interference pigments can be
obtained with an integrating sphere instrument when their optical
behavior and appearance are recognized.
The possible optical effects available to the artist or
designer are almost limitless and fascinating to imagine: pearlescent
pigments, used alone or in combination with transparent absorbing
pigments; iridescent (interference) pigments of specific hues, used alone
or in combination with transparent chromatic colorants. Changing the
substrate color with any of them is an additional variable that can be
added to the picture.
Although pearlescent and interference pigments may have
achieved up to now only limited usage in the works of artists and arti-
sans, one finds that pearlescent pigments have been included in tubes of
artists' colors for many years, with the original use as a pigment to simu-
late metals. For example, a pearlescent plus a little black makes a beauti-
ful silver and has been sold as "silver." With a little transparent red
178
Figure 5.16
The total reflectances of the same type of
red-blue (R-8) combination inference pig-
mented lacquer as in figure 5.15 are illus-
trated here. The curves were measured
with an integrating sphere instrument with
-8 incidence angle and the specular
reflectance included (total reflectance) .
When the pigment is applied over black,
only the interference blue color of 473 nm
dominant wavelength is apparent; all of
the transmitted light has been absorbed.
When it is applied over white, however,
the total reflectance includes not only the
transmitted color due to the carmine in
the Ti0
2
layer and the transmittance of
the interference pigment, but also the blue
interference specular color produced at
+8. The result is a magenta of 500c nm
dominant wavelength, which more closely
describes the color observed in semidiffuse
directional illumination, as would be seen
in commonly encountered illumination and
viewing conditions. The interplay of the
various types of reflections creates a
myriad of illusory colors.
Chapter 5
-v B G
y
0 R
100
R/ B anatase-mica combination interference
over white
90
pigment (sphere geometry measure)
80
70
<il 60
g
"'
u
c

u
50
"'


.....
c
"'
u
Q:j
40 CL
30
20
over black
10
400 450 500 550 600 650 700
Wavelength (nm)
oxide, a beautiful copper is achieved; with a little transparent yellow and
a little black, an interesting bronze. That these different optical effects
may be of interest to artists is attested to by the current offerings of tube
oil colors and acrylic colors by a well-known artists' materials supplier,
Daniel Smith (1992). His catalogue offers in oil colors three pearlescent
colors, fourteen iridescent colors, and seven interference colors. The iri-
descent colors and the interference colors, which contain coated mica
pigments, cover the visible spectrum. In the spring 1992 Daniel Smith
catalogue, an illustrated description of pearlescent and interference pig-
ments is presented by Weingrad (1992); color reprints are available for a
small fee from Daniel Smith, Incorporated, in Seattle. Weingrad presents
examples that an artist should experiment with to understand these pig-
ments. So, if pearlescent and interference pigments were not extensively
used by artists and artisans in the past, they may certainly appear in
future objects.
Color in Specular (Mirror-Type) Reflection 179
Throughout history, humans have thoroughly appreciated the
beauty of pearlescence and iridescence, using shells and pearls, and even
beetles and butterflies, to make jewelry and other objects for aesthetic
enjoyment. The attempt to use such materials for the artistic creation of
objects of beauty and adornment was a natural development. Because so
much of color in nature is a result of iridescence (interference), its mea-
surement and characterization may be of particular concern in the care of
collections in museums of natural history. Interference colors do not fade
in the same sense as dyes and pigments fade-by loss of the colorant
chemical compound. The deterioration that does affect the stability of
interference colors occurs as a result of minute changes in structure due,
for example, to shrinkage. Simon (1971) states that "even the most
infinitesimal alteration in these structures can lead to the colors' winking
out as though a switch had been turned off." Climate control would seem
to be critically important, then, in preserving the color of these objects.
It must not be forgotten that many of the colors observed in
nature's objects are not attributable to interference phenomena but are,
rather, caused by natural pigments and dyes, many of which are extremely
fugitive. Control of the intensity and spectral distribution of the illumina-
tion of exhibits of such materials is, therefore, vital to their color preser-
vation, in addition to the attention given to the control of climate.
Other Flake Pigments (Metallic Flakes)
Just as metals reflect light specularly (described in the first section of this
chapter), so too do flake pigments made from metals. When incorporated
into polymer films, they reflect like little mirrors.
For at least four thousand years, flake made from gold, the
so-called gold leaf, has been used for decoration. For most of this period,
the material was made by hand, with the gold hammered into very thin
sheets. Because of the high cost of gold leaf, development of a less-
expensive substitute was found in the form of a copper alloy such as
brass (copper alloyed primarily with zinc) or bronze (copper alloyed
with tin ). These substitutes were still hammered by hand, so the products
remained relatively expensive, though much less so than gold leaf.
Rogers, Greenawald, and Butters (1973), in their chapter in
the Pigment Handbook, tell an interesting historical anecdote concerning
the mechanization of the preparation of such substitutes. It seems that
Bessemer (of steel fame) discovered that gold bronze powder, which he
purchased around 1843-44 for his sister's book illustrations at what
seemed to him a ridiculously low price, on analysis turned out to be
brass. He put his inventive mind to work and developed, around 1845,
a mechanized process for making gold bronze powders. Never patented,
the secret process was kept within the family for many years. The same
process was used to make aluminum flakes until the early 1930s, accord-
ing to Rolles (1973). He also describes the old "secret" Bessemer ham-
mering process. Small, uniform pieces of metal were fed onto the steel
anvil of a stamping mill. A series of steel hammers was raised by action
180 Chapter 5
of a cam system and then allowed to fall onto the pieces on the anvil,
producing large flakes, called schrodes. The schrodes were then used as
the feedstock for smaller hammer mills to further foil out the material.
To prevent cold welding, lubricating oils and tallow fats were used. The
final step in the process was polishing the foil with rotating brushes.
Thus, the "secret" Bessemer process was simply the use of machines to
do the work formerly done by hand.
From the days when the imitation gold flakes first appeared,
the term bronze powder came to be used to describe any metallic flake
pigment, regardless of its metal composition. Today the term bronze
pigment may be preceded by the identification of the metal used, such
as "aluminum bronze" or "copper bronze."
Metallic flake pigments may be made from copper, brass,
bronze, or aluminum (the last was introduced at about the beginning
of the twentieth century) or from stainless steel (introduced around the
late 1940s). The brass and then the aluminum flakes, used primarily in
artists' materials when first introduced at the turn of the century, are
probably those most commonly encountered in addition to the real gold
leaf, which is still relatively expensive.
Stainless-steel flake pigments are used in coatings, primarily
for their very good durability and corrosion protection. Being of high
density, they do not "leaf," as do the other less-dense flake pigments,
but tend to form, as described by Hay (1973), a "multilayer structure
throughout the paint film." Because of their hardness, stainless-steel flake
pigments provide superior abrasion resistance in paint films, becoming
"brighter and smoother after several years of weathering" (Hay 1973).
Because of their chemical inertness, they provide superior corrosion resis-
tance in extremely antagonistic atmospheres. Also because of their inert-
ness, they may be combined with many types of additives to create
desired properties, or combined with other pigments to create desired
colors. They may also be formulated in almost any paint vehicle, oil
based or water based.
In contrast to what happens with the pearlescent and irides-
cent flake pigments described in the previous section, the change in color
with angle of view of paints pigmented with metallic flake pigments is
due simply to the change in the amount of light reflected by the more or
less horizontally oriented flakes. Thus, viewing at an angle near to the
specular angle, based on the angle of incidence used, results in the bright-
est color. As the viewing angle increases off the specular, the amount of
flake reflectance decreases, and the perceived color darkens.
The 50% cumulative (median) particle size of bronze
flake varies from 32 m for coarse particles to 6 m for fine particles
(Rogers, Greenawald, and Butters 1973); for aluminum flake, from
19 m for coarse particles to 7 m for the finest (Rolles 1973); and
for stainless steel, the median particle size is about 14 m (Hay 1973 ).
In contrast, the thickness of the flakes is much less-about 0.1-2.0 m
for aluminum flake, for example (Rolles 1973). The particle size of flake
pigments is within the range of a light microscope, making their study
reasonably simple.
Figure 5.17
Photomicrograph of a cross section of a
paint film pigmented with aluminum flakes
and phthalocyanine blue pigment. The
substrate (white section at the bottom) is
metal. The white flakes in the dark field
are aluminum flakes. The photograph illus-
trates how they are typically oriented
nearly parallel to the substrate interface.
Figure 5.18
Illustration of the orientation of metallic
flakes in a paint film, and the reflections
and interreflections that can occur at
an angle of incidence slightly off the
perpendicular.
Color in Specular (Mirror-Type) Reflection 181
The amount of metallic aluminum flake used in industrial
coatings is generally in the range of 2-3 % by weight of the total dry
film. Higher concentrations result in an increased loss of gloss of the
vehicle surface (Rolles 1973 ). Today overcoats of clear, transparent pig-
mented or unpigmented vehicle may be applied to the surface to mini-
mize this loss in gloss. This practice also enables the using of flakes with
the largest particle size while still retaining high gloss.
Figure 5 .17 is a photomicrograph of a cross section of a
film pigmented with aluminum flakes and phthalocyanine blue pigment
(Johnston 1967a) applied over a metallic substrate. It can be seen that
most of the flakes are oriented primarily parallel to the substrate, as
illustrated in the schematic drawing shown in figure 5.18.
Figure 5.19 illustrates the changes in reflectance with angles
of illumination and viewing for a film pigmented with aluminum flake.
The angles of incidence and viewing are described in degrees, the per-
pendicular to the surface being designated 0. The notation used in the
----
----
Metallic paint film
182
Figure 5.19
Variable-angle goniospectrophotometric
measurements of a paint film pigmented
only with aluminum flake pigment, identi-
fied by the composition description 2, 0, 0,
0 (the first number represents the weight
percent of aluminum flake, the second
number the percent of chromatic pigment,
the third number the percent of white pig-
ment, and the fourth number the percent
of black). The incident angle in degrees is
given first, negative if on the side of the
perpendicular opposite the viewi ng angle
(which is always positive), followed by a
slash, and then the viewing angle. The
number in parentheses is the number of
degrees off the specular angle. Note that
the measurements made with different
numbers of degrees off the specular-that
is, the refl ectance measurements made at
-30 /0 and at 0/30-are not identical,
as one might expect, because both are
measured at 30off the specular angle. At
least part of the explanation is the differ-
ences in the light incident (nearly colli-
mated) and the light viewed (subtending
an angle of about 15). The darkest color
is observed when the incident and viewing
angles are farthest from the specular
angle, that is, 0/75.
Chapter 5
-v B G
y
0 R
100
-30' /45' (15' )
90
80
-30' /10' (20' )
70
composition: 2,0,0,0
60
<!)
u
c:
.r:l
u
-30' /0' (30' ) <!)
""
50
<!)
>
0' /30' (30' )

<!)
a:
40
30
0' /45' (45')
20
0'/75' (75' )
10
400 450 500 550 600 650 700
Wavelength (nm)
descriptions of the geometrical spectral measurements is described in
chapter 1 in the discussion on goniospectrophotometry. The incident
angle is noted first, positive if on the same side of the perpendicular (0)
as the viewing angle, and negative if on the opposite side. This number
is foll owed by a slash and then the viewing angle, always positive. The
number of degrees off the specular angle is indicated in parentheses, as
shown in figures 5.19-5.23 and in table 5.3. The use of parentheses is
not standard, however; there is no method of notation generally accepted
at this time. Until there is, the notation described provides a convenient
shorthand description of the geometrical conditions of measurement.
The goniospectrophotometric measurements on metallic
paints illustrated in figure 5.19 and in subsequent figures (figs. 5.20-
5 .24) were made with the Trilac goniospectrophotometer manufactured
by Leres in France (see chap. 6). Measurements were made relative
to a pressed BaS0
4
white standard. It must be emphasized that the
Color in Specular (Mirror-Type) Reflection 183
Measurement geometry
Composition(%) Incident/ Deg. off
Figure Al, QM, W, Blk viewing angle specular angle x y
y
Comments
5.19 2, 0, 0, 0 -30 / 45 (15) 0.3096 0 .3169 94.74 Al alone
5.19 2,0, 0, 0 -30 /10 (20) 0.3091 0.3168 78.54 Al alone
5.19 2, 0, 0,0 -30 /0 (30) 0.3077 0.3153 56.38 Al alone
5.19 2, 0, 0, 0 0 /30 (30) 0.3076 0.3149 47.74 Al alone
5.19 2,0, 0, 0 0/45 (45) 0.3054 0.3127 30.79 Al alone
5.19 2, 0, 0, 0 0175 (75) 0.3030 0.3180 17.21 Al alone
5.20 2, 3, 0, 0 -30 /45 (15) 0.3109 0.2475 24.90 Al+QM
5.20 2, 3, 0, 0 -30/ 10 (20) 0.3113 0.2390 21.05 Al+QM
5.20 2,3,0,0 -3010 (30) 0.3149 0.2330 14.88 Al+QM
5.20 2, 3,0,0 0 /30 (30) 0.3183 0.2391 12.14 Al+QM
5.20 2, 3,0,0 0/45 (45) 0.3281 0.2354 7.90 Al+QM
5.20 2, 3, 0, 0 0175 (75) 0.3467 0.2326 4.34 Al+QM
5.21 2, 0, 0, 0.1 -30 / 45 (15) 0.3138 0.3209 72.00 Al+Blk
5.21 2, 0, 0, 0.1 -30 /10 (20) 0.3136 0.3211 60.35 Al+Blk
5.21 2, 0, 0, 0.1 -3010 (30) 0.3132 0.3206 41.97 Al+Blk
5.21 2, 0, 0, 0.1 0 /30 (30) 0.3132 0.3203 35.84 Al+Blk
5.21 2, 0, 0, 0.1 0/45 (45) 0.3124 0.3196 22.32 Al+Blk
5.21 2, 0, 0, 0.1 0175 (75) 0.3125 0.3196 11.68 Al+Blk
5.22 2, 0, 2, 0 -30 /45 (15) 0.3065 0.3124 55.40 Al+W
5.22 2, 0, 2, 0 -30/ 10 (20) 0.3058 0.3115 47.95 Al+W
5.22 2, 0, 2, 0 -3010 (30) 0.3043 0.3102 38.84 Al+W
5.22 2, 0, 2, 0 0 /30 (30) 0.3031 0.3098 35.13 Al+W
5.22 2,0, 2,0 0/45 (45) 0.3011 0.3090 29.40 Al+W
5.22 2, 0, 2, 0 0 /75 (75) 0.2986 0.3079 23.80 Al+W
5.23 2, 0, 2, 0.1 -30 /45 (15) 0.3136 0.3186 45.29 Al+W+Blk
5.23 2, 0, 2, 0.1 -30/ 10 (20) 0.3115 0.3174 34.04 Al+W+Blk
5.23 2, 0, 2, 0.1 -3010 (30) 0.3102 0.3166 26.06 Al+W+Blk
5.23 2, 0, 2, 0.1 0 /30 (30) 0.3117 0.3171 27.80 Al+W+Blk
5.23 2, 0, 2, 0.1 0/45 (45) 0.3109 0.3178 21.83 Al+W+Blk
5.23 2,0,2,0.1 0175 (75) 0.3098 0.3178 16.14 Al+W+Blk
aptgment weight concentration in the dried film. Al =aluminum flake; QM= Quinacridone
Magenta; Blk =carbon black; W = rutile Ti0
2
white.
Table 5.3
CIE notation (illuminant C, 1931 standard
observer) for metallic paints.
reflectances measured are relative reflectances, valid only for the geo-
metric conditions used in the measurements. It must also be remembered
that reflectances (strictly speaking, reflectance factors) as conventionally
measured are the ratio of the intensity of the light reflected from the
sample, I, to that reflected from the reference white standard, 1
0
-that
is, R = I/1
0
. In the case of variable-angle measurements, the same is true.
However, if measurements are made at the specular angle on flake pig-
ments or on highly reflecting metals, the results will be well above the
near 100% reflectance of the white reference standard. For example,
the reflectances of the interference flake pigments at the specular angle
describing the interference color, as reported by Greenstein (1973a) and
in the previous section, required the use of a neutral-density filter in the
sample beam, generally one of 10% transmission. Hence, Greenstein's
use of the term "relative reflectance" and the author's use of the term
correspond here. (For the measurements reported here on the metallic
flake pigments, no neutral-density filter was used, because the measure-
ments were made at angles well removed from the specular angle.)
184
Figure 5.20
Variable-angle goniospectrophotometric
measurements of a paint film made with
the same aluminum flake pigment in the
same amount as in figure 5.19, but with
3 % chromatic pigment (Quinacridone
Magenta) added. The composition is
described as 2, 3, 0, 0 (for explanation,
see fig. 5.19). The measurement angles
are the same as in figure 5.19.
Chapter 5
y 0 -----R----
90
composition: 2,3,0,0
80
70

60
:::l
:;:
:s
Q)
u
c
l'l
50 u
Q)
;:;:::
!'::'
Q)
>

40 a:;
er::
30
20
10
400 450 500 550 600 650 700
Wavelength (nm)
From an examination of the measurements for aluminum
flake alone (without a colored pigment or an inert present), illustrated in
figure 5.19, the following observations can be made: (1) the closer the
viewing angle is to the specular angle, the higher is the reflectance of
the aluminum flake paint; and (2) the magnitude of the reflectance at a
specified off-specular viewing angle (30, as illustrated) varies somewhat
with the direction of the incident angle. (The optics of the measuring
instrument are not reversible.) The CIE notations (illuminant C, 2 stan-
dard observer) for all of the subsequent illustrations of metallic finishes
are described in table 5.3.
It should be noted that the CIE color differences calculated
for "relative reflectance" measurements do not have the same meaning as
those calculated from "absolute reflectance" data for nondirectional
samples measured on conventional color-measuring instruments.
Corrections can be made for the absorption of the neutral-density filters
Figure 5.21
Variable-angle goniospectrophotometric
measurements of a paint film containing
the same aluminum flake pigment as in
figures 5.19 and 5.20, but containing only
0.1 % carbon black in addition to the alu-
minum flake. The same angular conditions
as in figures 5.19 and 5.20 were used.
Note that reflectances at all angles of illu-
mination and viewing are decreased, as
would be expected.
Color in Specular (Mirror-Type) Reflection
100
90
80
70
60
Q)
u
c

u
Q)
""
50
Q)
>
:p
n;
OJ
er:
40
30
20
10
400 450 500
y 0
composition: 2,0,0,0.1
-30/45(15)
-30/10(20)
-30/0(30)
0130(30)
0/45(15)
0/75(75)
550
Wavelength (nm)
600
185
-----R----
650 700
used, however. Even without such corrections-but assuming that the
neutral-density filter transmission is very flat-the CIE chromaticity coor-
dinates, x and y, of the metallic finishes still provide useful information
for comparison to those of similar samples measured in a similar fashion.
In summary, it should be remembered that the relative gonio measure-
ments are not absolute in the normal sense, as are measurements made
with conventional color-measuring instruments intended to measure dif-
fuse or total reflectance. Note also that secondary standards for the dif-
fusing reference white, such as white glass or porcelain, should not be
used unless they have been calibrated for the particular angular condi-
tions used or are fully described for comparison purposes.
The effect on the reflectance curve resulting from the addition
of a relatively transparent absorbing pigment (Quinacridone Magenta) to
the same amount of the same aluminum flake in figure 5.19 is illustrated
in figure 5 .20. The angles of illumination and view are the same as in
186
Figure 5.22
Variable-angle goniospectrophotometric
measurements of a paint film containing
the same aluminum flake in the same
amount as in figure 5.19, but also contain-
ing 2 % rutile Ti0
2
. Reflectances at near-
specular angles are decreased, but at lower
angles of view relative to the incident
angles, reflectances are increased.
Chapter 5
y 0 -----R----
90
80
composition: 2,0,2,0
70
60
-30' /45' (15)
"'
u
c
.r:l
u
"' '+=
50
-30/10(20)
"'
>

Q)
er::
-30' /0' (30')
40
0' /30' (30' )
30
0' /45' (15' )
20
0'/75' (75' )
10
400 450 500 550 600 650 700
Wavelength (nm)
figure 5.19. It can be seen that the addition of a transparent absorbing
pigment decreases the reflectance at all angles and that the three charac-
teristic absorption bands of the Quinacridone Magenta remain obvious.
(Compare with curves for the closely related Quinacridone Red illus-
trated in fig. A.16.)
The effect of the addition of carbon black, a transparent,
nonselective absorber, is illustrated in figure 5 .21. Again the reflectance is
lowered (absorption is increased) at all angles of illumination and view.
This is as would be expected from Kubeika-Munk theory, described in
chapter 4, and from the Beer-Bouguer equation.
However, the effects due to the other constant in the
Kubeika-Munk equation-the scattering coefficient, 5-are far more
complex, because light scattering involves so many factors and often
gives rise to confusing and unpredictable results. It is particularly puz-
zling when small amounts of a scattering pigment are added to a metallic
Figure 5.23
Variable-angle goniospectrophotometric
measurements of a paint film containing
the same aluminum flake in the same
amount as in figure 5.19, the same
amount of black as in figure 5.21, and
the same amount of white as in figure
5.22. The effect of the combination of
white Ti0
2
with the black is an additive
one. At the near-normal viewing angle-
for example, 0/30-the color is darker
than it is when either the black alone, fig-
ure 5.21, or the white alone, figure 5.22,
is added. However, at low angles of
viewing (0/75), the color is lighter than
it is when only black is added, figure 5.21,
but darker than it is when only white is
added, figure 5.22. In this case, the light-
ening effect of white at low angles of
viewing is stronger than the darkening
effect of black.
Color in Specular (Mirror-Type) Reflection 187
flake paint film. The addition of a small amount of Ti0
2
, representative
of a scattering pigment, to an otherwise clear paint made with aluminum
flake pigment, results in a darkening-as though a black pigment had
been added-when the paint is observed near to the specular angle; the
addition does not increase the reflectance, as would be expected from
Kubeika-Munk theory. At a greater distance from the specular angle,
however, the reflectance increases, distinguishing the sample from a film
made with black and aluminum flake. This effect is illustrated in figure
5.22 and is summarized in CIE terms in tabl e 5.3.
Figure 5 .23 and table 5 .3 illustrate the goniospectrophoto-
metric measurements and CIE coordinates of a sample containing both
black and white with the same amount of aluminum flake. The effects of
each of the pigments, the absorber (black) and the scatterer (Ti0
2
) , are
still apparent in the mixture. Figure 5 .24 shows a graph of the luminous
reflectance, Y, versus the various degrees of view off the specular angle.
y 0 -----R----
90
80 composition: 2,0,2,0.1
70
60
.,
u
c
.rl
u
.,
;:;:::
1:: 50
.,
-30/45(15)
,;::

.,
er::
40
-30/ 1 Q (20)
30
0130(30)
-3010(30)
20
0/45(15)
0/75(75)
10
400 450 500 550 600 650 700
Wavelength (nm)
188
Figure 5.24
Relative luminous reflectance, Y. plotted
against the degrees of the viewing angle
off the specular angle. The addition of K
refers to the addition of the black absorb-
ing pigment, and S to the addition of the
white scattering pigment. Note that the
reflectances measured at different incident
angles are different, even when the mea-
surement is made at the same number of
degrees off the specular angle, 30, as
illustrated here. Obviously, with this
instrument the optics are not reversible.
Chapter 5
90
80
70
>-
~ 60
c
~
u
Q.J
;:;:::
~
~ 50
0
c
E
::l
Q.J
>
~ 40
Q:;
er::
30
20
10
\
4
\
addition of K and S to metallic paint films
(O' and -30' incidence)
Al alone (2,0,0,0)
- - -b - - Al+ black (2,0,0,0.1)
-O- Al+ white (2,0,2,0)
--0- Al+ W + Blk (2,0,2,0.1)
........... --6..
0 ' - - ~ ~ - - ' - ~ ~ ~ - - ' - ~ ~ ~ ' - - ~ ~ - - - - ' - - ~ ~ - - ' - ~ ~ ~ - - ~ ~ ~
(10' ) (20' ) (30) (40) (50) (60' ) (70' ) (80' )
Viewing angle (degrees off specular angle)
Thus, the addition of a small amount of a scattering pigment (white)
lowers the reflectance near the specular angle but increases the
reflectance well off the specular angle of view, even when an absorbing
pigment is present. Further increasing the concentration of a scattering
pigment, however, results in the more expected behavior of the addition
of a scattering pigment-that is, lightening of the color as observed at
angles far from the specular angle. But when the concentration of scatter-
ing pigment is increased, the specular reflectance of the metallic flake pig-
ment is obscured, and the metallic appearance is decreased.
It should be emphasized that such darkening near the specu-
lar angle owing to the addition of a small amount of a scattering pigment
occurs whenever any scattering pigment (type I or type III, described in
chap. 3) is added, regardless of its hue. Thus, the addition of a small
amount of a scattering yellow oxide pigment to a metallic paint film will
Color in Specular (Mirror-Type) Reflection 189
have a similar effect: darkening of the so-called face color-that is, the
color near to the specular angle-and lightening of the so-called flop
color, observed well off the specular angle.
The extent of the use of metallic flake pigments in materials
of interest in museum collections is not known. However, they may
occur in artists' works and in commercial products made since the turn
of the last century. Moreover, they may be used increasingly in the future
due to their inclusion in artists' tube oil paints by many colorant manu-
facturers (Smith 1992). One recent catalogue lists eight watercolors
varying from silver to gold to copper. Recognizing the possibility of their
use is being forewarned.
Chapter 6
Special Topics
The discussion in this chapter involves the description of material charac-
teristics that affect color and appearance but that cannot be attributed to
the pigment absorption and scattering described by the Kubeika-Munk
theory (chap. 4) or to the special properties of metals and flakes (chap. 5).
The characteristics to be considered include surface structure (such as
glossiness and matteness), as well as fluorescence, extender pigments, and
internal structure, such as microvoids. All of these factors may affect the
perceived color and the difference in color between similar objects. It is
often difficult to ascertain the cause of observed differences between
objects as a result of such factors or to visualize the changes they may
induce in color.
Surface Reflection
Objects can exhibit surface reflection that varies between a high gloss
and a flat or matte (diffusely reflecting) appearance. Between these two
extremes, materials may exhibit a myriad of textures, from "micro"
roughness (beyond visual resolution), which is often categorized as semi-
gloss, to various obvious "macro" surface irregularities. In addition to
those due to textural roughness, color variations may be encountered
that are much more subtle. Mottled color variations will not be consid-
ered here (see chaps. 2 and 8 ).
Refractive index differences: The cause of surface reflection
That surface reflection always occurs when incident li ght encounters a
material of different refractive index has been emphasized previously.
For dielectric (nonelectricall y conducting) materials, the amount of the
surface reflection is described by equation 4.3 and illustrated in figure
4.2. The amount of the incident light reflected depends on the ratio of
the refractive indices of the two materi als at the interface. In the case of
the upper surface reflection, the incident light passes from air, with
refractive index n1 (in eq. 4.3, this is set equal to 1.000), to the surface
material, of refractive index n 2 It should be emphasized, however, that
interfaces between materials can occur throughout the body of mixtures,
not just at the air-material interface. Indeed, the opacity exhibited by
Special Topics 191
pigments is dependent on surface reflection at the interfaces of the pig-
ment particles with the vehicle in which they are suspended. Although
most examples presented here refer to pigmented materials such as
paints, surface reflection always occurs from objects or we would not be
able to see them.
The refractive index of a material changes with wavelength.
Commonly cited, so-called characteristic refractive indices of materials
are assumed to be those measured at 589 nm, the median wavelength of
the sodium D lines, unless specifically designated otherwise. In the case
of nonmetallic materials, the refractive index and, therefore, the surface
reflection, changes so slightly with wavelength that, for practical pur-
poses, it can be treated as a constant. (An exception occurs in the case of
bronzing described in chap. 5.)
When glossy samples are viewed, the specular reflection from
the surface is excluded from the visual evaluation of color. To illustrate
the effect of this exclusion on the color, consider an off-white glossy
sample that reflects 80% diffusely and 4 % specularly. The surface
reflection of 4% in this case represents 4/84, or 4.8%, of the total
reflected light. On a dark glossy material of the same surface composi-
tion that has a total reflectance (diffuse plus specular) of 10%, the same
4 % now represents 4110, or 40%, of the reflected light. Thus, small dif-
ferences in the surface reflection characteristics are far more significant in
dark colors than in light colors, a fact well appreciated by artists and
conservators. In the case of high-chroma colors (where in one wavelength
region the total reflectance is very low, such as the 10% illustrated
above, and in another region of the visible spectrum it is very hi gh, such
as the 84% for the off-white described above), differences in surface
reflection alter the purity or saturation of the perceived color as well as
its lightness, but they do not alter its dominant wavelength or hue. When
the surface is not glossy, the diffuse surface reflectance is essentially the
color of the light source and is additively mixed with the internal diffuse
reflectance, which lowers the perceived chroma.
In what follows, the microstructure of surfaces is considered
first . When matte s urfaces are compared, the structure of this surface is
one of the most insidious aspects affecting their appearance-insidious
because it is difficult both to identify precisely and to adjust in order to
create an identical appearance in two objects. As has been emphasized
previously (chap. 4 ), the nature of the first-surface reflection plays an
important role in determining the perceived color that occurs as a result
of the selective reflection of the colorants within the object. An extensive,
annotated bibliography, Matte Paint, has been published by Hansen,
Walston, and Bishop (1993 ).
In this initial discussion of micro surface reflection differ-
ences, the assumption is made that the sample surface exhibits rotational
symmetry: at each set of incident and viewing angles, the sample may be
rotated in its own plane through 360 without exhi biting a significant
change in the measured reflectance. The assumption is also made that the
pigmentation is the same in the two members of each sample pair.
192
Figure 6.1
Variable-angle reflectances measured on
matte paints, a violet and a purple, both
pigmented with carbazole dioxazine violet
and rutile Ti0
2
. The measurements were
made relative to Ba50
4
white standard.
Note that the reflectance increases as the
angle of viewing increases away from the
incident angle, which is also the specular
angle; reflectance continues to increase to
75 for both the purple and the violet.
Chapter 6
Matte surfaces that exhibit geometric metamerism
As was pointed out in chapter 2, slight differences in surface structure
between pairs of samples of the same pigmentation can lead to geometric
metamerism-that is, the pair can appear to be identical in color at one
set of illumination and viewing angles but no longer appear to be the
same when the viewing or the illumination angle is changed. Figures 6.1
and 6.2 illustrate the variable-angle reflectance measurements, relative to
pressed BaS04 reference white, of two sets of two differently pigmented
paint panels-one a violet (light purple), the other a purple-exhibiting
surface reflections that seem to be similarly diffuse. The violet and purple
samples numbered 1 in figure 6.1 have the same matte surface character-
istics. The two samples labeled 2 in figure 6.2 also have the same surface
characteristics, with the exception that they also have been coated with a
vehicle containing a silica flatting agent. Both appear to be matte sur-
faces. It can be seen that the samples in figure 6.1 exhibit increasing
Q)
v
c

u
Q)
;:;:::

Q)
>
:;::;
"'
Q)
er:
- v--11-c--- B -------- G
y 0
90
80
70
60
50
40
30
20
10
400 450 500
matte paints: samples #1
i = 0
550
Wavelength (nm)
600
----- R----
650 700
Figure 6.2
Variable-angle reflectances measured on
the violet and purple matte paint panels
coated with a thin application of vehicle
containing a small amount of silica flat-
ting agent. The appearance is still that
of a matte paint. Note that the highest
reflectance for both the violet and the
purple is that made at 30; reflectance
decreases as the viewing angle increases
away from the specular angle. In contrast,
the maximum reflectance for the uncoated
samples in figure 6.1 is farthest away from
the specular angle.
Special Topics 193
- v---11- c--- B - ------ G
y 0
----- R----
100
90
80
70
60
Q)
u
c
.s
u
Q)
""
l': 50
Q)
>

Qj
er:
40
30
20
10
400 450 500
matte paints: samples #2
i = 0
550
Wavelength (nm)
600 650 700
reflectance as the viewing angle increases away from the specular angle,
0. On figure 6.2, however, the samples do the opposite-that is, their
reflectances decrease as the viewing angle increases away from the specu-
lar angle. The samples were made with carbazole dioxazine violet pig-
ment in mixture with white (see fig. A.25).
Table 6.1 gives the CIE tristimulus data relative to pressed
BaS04 reference white for the variable-angle measurements shown in
figures 6.1 and 6.2.
The same Trilac spectrophotometer used for the variable-
angle measurements was also used in the sphere mode, with the same
pressed BaS04 white standard. CIE data calculated from these measure-
ments are tabulated in table 6.2.
Comparison of the color differences, .1.E''. ab, calculated
from the sphere-mode measurements {table 6.2) and from the variable-
angle measurements {table 6.1 ) illustrates the lack of sensitivity of the
194 Chapter 6
Sample
;o/vo
x
y
z x y L* L*
2
/L*
1 LiE*ab (Std. = 1)
Violet 1 0130 39.08 35.43 61.18 0.2880 0.2611 66.09
Violet 2 0130 40.27 36.72 60.92 0.2920 0.2662 67.07 1.015 2.2
Purple 1 0130 15.38 11.93 31.72 0.2606 0.2021 41 .1 1
Purple 2 0130 16.79 13.61 33.15 0.2642 0.2142 43.67 1.062 4.7
Violet 1 0145 40.06 36.22 62.08 0.2895 0.2618 66.69
Violet 2 0145 38.55 34.84 59.00 0.2912 0.2613 65.63 0.984 1.6
Purple 1 0145 15.61 12.04 32.33 0.2602 0.2008 41 .28
Purple 2 0145 14.58 11.10 30.35 0.2603 0.1980 39.75 0.963 1.4
Violet 1 0175 43.72 39.88 66.92 0.2904 0.2649 69.39
Violet 2 0175 38.40 34.67 58.62 0.2916 0.2633 65.49 0.944 4.0
Purple 1 0175 17.26 13.52 35.64 0.2599 0.2035 43.54
Purple 2 0175 14.08 10.55 30.07 0.2575 0.1928 38.82 0.892 5.2
i
0
= incident angle; v
0
= viewing angle; 0 = perpendicular to the surface
1 = sample 1 in figure 6.1 ; 2 = sample 2 in figure 6.2
Table 6.1
Variable-angle measurements on violet
and purple pairs of matte paints (CIE
coordinates; illuminant C; 1931 standard
observer; color difference equation,
eq. 2.4) .
Table 6.2
Violet and purple matte paint pairs mea-
sured with integrating sphere geometry
(CIE coordinates; illuminant C; 1931 stan-
dard observer; color difference equation,
eq. 2.4).
Sample Geometry x
Violet 1 SCE 40.31
Violet 2 SCE 39.76
Purple 1 SCE 16.28
Purple 2 SCE 16.44
Violet 1 SCI 39.74
Violet 2 SCI 39.73
Purple 1 SCI 16.06
Purple 2 SCI 16.45
conventional sphere measurement to the subtle surface angular reflec-
tance differences in these samples. Attempts to reconcile such mea-
surements with visual evaluations can be frustrating indeed.
An additional comment can be made on the sphere-mode
measurements: Note in table 6.2 that the total reflectance measurements,
SCI, are lower on most of the matte number 1 samples (uncoated) than
the diffuse, SCE, measurements. This phenomenon is encountered often
on very matte surfaces made with Ti0
2
This pigment has a significantly
higher refractive index than the pressed BaS0
4
standard and hence
higher surface reflectance (see Kortum 1969). In addition, pressed BaS0
4
has a slight sheen at angles approaching grazing angle that is not encoun-
tered with the most diffuse Ti0
2
pigmented matte paints. The effect of
Ti0
2
on surface characteristics has been discussed by Braun (1991).
The relationships between the angular reflectances of the two
types of surfaces described above for samples 1 and 2 can be seen more
vividly by plotting their lightness, L'', against the angle of view (the same
as the angle off the specular angle because the incident angle is 0 ), as
shown in figure 6.3a. At about 38 to 40, the lightness values cross, as
clearly shown by the lightness ratios plotted in figure 6.3b. This measured
reversal with angle of viewing is in agreement with visual observation. (It
is interesting to note that the angle at which many geometric metamers
LiE *ab
y
z x y L* L*
2
1L*
1
(Std.= 1)
37.1 7 62.09 0.2888 0.2663 67.41
36.66 59.99 0.2914 0.2687 67.02 0.994 1.21
12.98 34.07 0.2570 0.2050 42.74
13.29 33 03 0.2607 0.2108 43.20 1.011 1.01
36.59 61 .54 0.2882 0.2654 66.97
36.60 60.26 0.2908 0.2679 66.98 1.000 1.14
12.82 33.63 0.2569 0.2050 42.49
13.34 33.25 0.2609 0.2116 43.27 1.018 1.02
' SCE =specular component (of the reflectance) excluded; SCI =specular component (of the reflectance) included
1 = sample 1 in figure 6.1 , 2 = sample 2 in figure 6.2
(a)
70

60
L
50
40
20 40 60
Viewing angle
Figure 6.3
(a) Values of L * for the variable-angle
measurements illustrated in figures 6.1 and
6.2, plotted against the viewing angle
when the incident angle is 0. Note that
the two types of surfaces match at about
40 viewing. (b) The ratio of the L * mea-
surements of the violet and purples, based
on the data shown in (a).
#1
#2
80
Special Topics 195
violet (b)
1.1
#2/#1
L*2
1.0
L*1
violet
0.9
purple
purpl e
0.8
20 40 60 80
Viewing angle
match is at about 45 when the incident illumination angle is near nor-
mal; see the discussion on flake pigments in chap. 5, for example.)
These samples were prepared intentionally to illustrate (1) the
phenomenon of geometric metamerism commonly encountered with
matte-surface materials, and (2) the effects of changes in micro surface
structure on the perceived color of matte surfaces. The two colors were
also chosen to illustrate that surface reflection differences are visually
more important as the colors become darker. Compare the color differ-
ences, of the violets relative to those of the darker purples; the lat-
ter exhibit larger color difference when measured at 0130 and at 0/75.
The color differences are similar in magnitude at the angle close to the
match angle, 0145. Samples numbered 1 were used as the standard.
In this illustration, sample 2 is the same as sample 1, but the
difference between them arises because sample 2 was sprayed with a thin
layer of clear vehicle that also contained a small amount of a silica
flatting agent. Thus, the samples exemplify a type of problem often
encountered when one attempts to consolidate powdery or matte surfaces
of museum objects in order to preserve them. An appreciable application
of consolidating resin changes the optics of the surface reflection and
simultaneously changes the color perceived at various angles.
Figure 6.4 helps explain, in a somewhat simplistic manner,
the differences in surfaces that can occur when a clear film of vehicle is
applied over a very diffuse surface. Illustration (a) represents the diffuse
surface; illustration (b) shows an excellent consolidant application in
which the applied vehicle penetrates into the interstices (air spaces)
around the pigment on the surface but does not form a film on the upper
surface; illustration (c) shows what occurs at the extreme case, in which
the applied vehicle does not penetrate into the interstices but stays on
top of the surface, forming a smooth, shiny film. A further difficulty can
contribute to a less-than-ideal penetration: if the resin solution does not
"wet"-that is, it does not adhere to the dry pigment on the surface-the
resin can form a film that has poor adhesion to the surface.
196
Figure 6.4
(a) Typical matte paint with pigment par-
ticles protruding through the surface,
causing scattering of the light. (b) Adding
a small amount of a vehicle applied to the
matte paint fills part of the air interstices
between the pigment particles. The result
is an increase in the specular reflectance
when the angles of incidence and viewing
are near to the perpendicular, and a dark-
ening when the angle of incidence is still
near to the perpendicular but the view-
ing angle is closer to the horizontal.
(c) Vehicle is applied over the matte sur-
face but does not penetrate and replace
the air in the pigment interstices. A film
is formed over the surface, gloss is
increased, and the color is thus darkened.
Although the surface has become glossy,
the entrapped air in the interstices scatters
light and partially offsets the darkening.
This effect is often the result of inade-
quate wetting of the surface, which also
leads to poor adhesion to the surface.
Chapter 6
(a) (b)
i
Wli:i
-
vehicle
0
pigment
(c)
unshaded = air
The film illustrated by (c) in figure 6.4 exhibits some specular
gloss, although the color is not as dark as would be expected if there
were no air pockets trapped under the film to scatter the light. Even with
the rather ideal coating application illustrated in (b), at near-normal illu-
mination and viewing, there is a slight increase in surface reflectance at
the angle nearest the specular angle. Such considerations imply that there
may be no perfect technique to consolidate powdery surfaces without
some change, however slight, in the perceived color.
Hansen, Lowinger, and Sadoff (1993) have developed a tech-
nique for achieving good penetration of resin solution into powdery sur-
faces such as are found on ethnographic objects. This technique involves
spraying the object and then keeping it in a closed container filled with
solvent vapor. This procedure allows the resin solution time to penetrate
into the interstices and replace the air around the protruding particles
before the solvent evaporates and a film forms. When the objects are of
complex shape or color, or both, the problem of surface alteration of this
type may be less acute than when the matte surface is perfectly planar
and consists of large areas of a uniform single color. In the latter situa-
tion every slight change is obvious.
Matte surfaces are not the only type that exhibit geometric
metamerism. Almost any type of surface, short of those with high gloss,
may exhibit the phenomenon. Intermediate-gloss surfaces present many
possibilities for subtle surface-structure differences.
Surface effects on low-chroma and high-chroma colors
The effects of surface differences on a high-chroma color relative to the
same differences on a neutral color can be illustrated by examining the
reflectance measurements made on two samples of gel-coated, glass-fiber
laminates, one a high-chroma red and the other a neutral gray. The sur-
face characteristics of gel coats, widely used to make glass-fiber laminate
boats, are determined by the surface characteristics of the mold employed
in their construction and are not attributable to differences in composi-
tion. (Gel coats contain no solvent; the liquids are polymerized in the
curing process. ) If the interior of the mold is very smooth and shiny, the
surface cast against it will be glossy; if the mold has a textured surface,
Figure 6.5
Spectral reflectance measurements of three
red plastic, gel-coated, glass-fiber lami-
nates exhibiting different surface charac-
teristics. Specular reflectances were
excluded. The panels were pigmented
with cadmium red medium.
Special Topics 197
so will the object cast in it. The amount of pigment in the plastic gel coat
is small relative to that used in paints-often less than 10%-and it does
not affect the surfaces.
Three types of mold surfaces were used for each of the red
and gray gel coats: a very high gloss; a matte; and a highly textured,
wrinkle effect. Figures 6.5 and 6.6 show the reflectance curves of these
six samples as measured with the specular component of the reflectance
excluded (SCE)-that is, diffuse reflectance. The two glossy samples
appear darker than the corresponding matte and textured samples. The
red glossy sample is obviously higher in chroma, or saturation, than
either of the other two reds. Figures 6. 7 and 6.8 depict the reflectances
measured with the specular component of the reflectance included
(SCI)-that is, total reflectance. All of the measurements were made with
the Trilac spectrophotometer operated in the sphere mode relative to
pressed BaS04 reference.
90
80
70
-;;; 60
g
Q)
u
c
1l
u
50
Q)
;:;:::


c
Q)

Q)
40 o._
30
20
10
400
y 0
red plastic gel coats (measured SCE)
450
matte
textured
glossy
500 550
Wavelength (nm)
600
i.---- R ----
650 700
198
Figure 6.6
Spectral reflectance measurements of three
gray plastic, gel-coated, glass-fiber lami-
nates exhibiting the same three surface
characteristics as the red panels in figure
6.5. Specular reflectances were excluded.
The panels were pigmented with rutile
Ti0
2
and black.
Chapter 6
90
80
70
<il 60
g
OJ
u
c
1l
u
50
OJ
""

....,
c
OJ
u
.,
40
"-
30
20
10
400
y 0
gray plastic gel coats (measured SCE)
450
matte
textured
glossy
500 550 600
Wavelength (nm)
--- -- R----
650 700
Figure 6.9 illustrates schematically the angular reflectances for
two different surfaces. The glossy surface reflectance illustrated in (a) has
a sharp, spiked specular reflectance, marked S in the figure, superposed
on the diffuse reflectance, D. When gloss decreases, the width of the spec-
ular spike increases, and the height of the spike decreases, as illustrated
in (c). One type of integrating sphere, discussed in chapter 1, is illustrated
schematically in (b) and (d). The illustrated technique using the spheres is
often used with spectrophotometers for measuring separately t he total
and diffuse reflectances. When the insert in the sphere wall is the same
white as the sphere interior wall, all of the light reflected by the sample is
included in the measurement (SCI); when the insert is black or is a light
trap, then the specular reflectance, S, is absorbed, and only the diffuse
reflectance (SCE) is measured. For the high-gloss sample, diagrammed in
figure 6.9 as (a), the specular reflectance, S, can be completely absorbed
by the black insert and eliminated from the measurement, as illustrated in
Figure 6.7
Spectral reflectances of the three red plas-
tic, gel-coated, glass-fiber laminates with
the different surface characteristics illus-
trated in figure 6.5, but measured with the
specular reflectance included, SCI (total
reflectance). Note that the reflectances are
very close.
Special Topi cs
90
80
70

60
g
Q)
u
c
_:g
u
50
Q)
""'

....,
c
Q)

Q)
40 a..
30
20
10
400
y 0
red plastic gel coats (measured SCI)
450
matte
textured
glossy
500 550
Wavelength (nm)
600
199
-----R----
.. ---
=---
650 700
(b). When the gloss is decreased, as illustrated in figure 6.9 as (c), the sur-
face reflectance at the specular angle can no longer be completely elimi-
nated by the black insert and is partially included in the measurement of
the diffuse reflectance, as illustrated in (d).
Other designs for including or excluding the specular surface
reflection on integrating-sphere spectrophotometers have been used, but
basically all give similar results.
Table 6.3 gives a summary of the tristimulus values and chro-
maticity coordinates (1931 2 standard observer, illuminant C) of the red
and gray plastic gel coats whose spectral reflectance curves are illustrated
in figures 6.5-6.8. Color differences of the matte and textured samples
were calculated relative to the corresponding glossy samples. Note that
the luminous reflectances, Y, of the two colors were not identical: for the
red (measured SCE), Y = 8.83; for the gray (measured SCE), Y = 6.87.
Nonetheless, they were not in an entirely different range.
200
Figure 6.8
Spectral reflectances of the three gray
plastic, gel-coated, glass-fiber laminates
with the different surface characteristics
illustrated in figure 6.6, but measured
with the specular reflectance included,
SCI (total reflectance). Note that these
reflectances are also very close.
Chapter 6
-----R----
90
80
70

60
g
<!)
u
c

u
50 <!)
<+=

_.._,
c
<!)

<!)
40
CL
30
20
10
400
gray plastic gel coats (measured SCI)
450
matte
textured
glossy
500 550
Wavelength (nm)
600 650
Inspection of the curves and the calculations measured
700
with the specular reflectance excluded (SCE) show the effect of surface
reflectance dissimilarities on the perceived color of the high-chroma red
samples relative to the same surface differences in the achromatic grays.
The matte red plastic panel and the glossy red panel are illustrated in
color plate 4. For the red panel, the color difference of the matte sam-
ples, M, relative to the glossy samples, G, is chiefly due to the chroma-
ticity difference, t-.C = 19.0, with a smaller effect due to the lightness
difference, t-..L ,,. = 6. 7. On the other hand, for the achromatic gray panel
with the corresponding surfaces, the chromaticity difference, t-..C, is very
small on the matte sample (equal to 0.43), whereas the total color differ-
ence, t-..E''. = 8.0, is almost completely due to the lightness difference,
t-..L ,,. = 8.0. Note also that the total color difference is much higher for
the red samples, t-..E'' = 20, versus t-..E'' = 8.0 for the corresponding
achromatic gray samples.
Figure 6.9
Angular distribution of the specular, S, and
diffuse, D, reflectances of a high-gloss sur-
face (a) and a semimatte surface (c). 1
0
refers to the incident light ray, near to the
perpendicular, or normal. Note that for
the semimatte surface, the specular peak,
S, is broadened and decreased in height.
Diagram (b) shows how the specularly
reflected light of a glossy surface is mea-
sured with an instrument that uses an inte-
grating sphere with the angle of incidence
near to the normal. When the sphere
insert at the angle equal to and opposite
the incident angle is black or a light trap,
the specular reflectance is absorbed, and
only the diffusely reflected light is reflected
onto the white sphere wall and recorded
by the detector. See the reflectance curves
Special Topics
D
s
glossy surface
(a)
sphere
201
D D D
semimatte surface
(c)
sphere
ref.
of the red and the gray glossy, plastic lami-
1
o
nates measured with the specular compo-
nent of the reflectance excluded (SCE) in
figures 6.5 and 6.6. Diagram (d) shows
how the specularly reflected light is mea-
sured for the semimatte surface with an
integrating sphere instrument. When a
black insert is placed in the sphere wall at
the angle equal to and opposite the inci-
dent angle, only a portion of the specular
reflectance is excluded. When a diffuse
white insert that matches the sphere wall
is placed in the sphere, all of the reflected
light is included in the measurement-that
is, the total reflectance is measured. See
the curves for the SCI measurements of
the red and gray plastic panels in figures
6.7 and 6.8, illustrating that they are the
same when all of the light is measured.
(Small differences may be due to wear and
tear on the panels; they were rather old
when these measurements were made.)
spl.
glossy surface semimatte surface
(b) (d)
Although it is not readily apparent from these data, the chro-
maticity difference, L'lC, which is so high on the red matte and textured
samples relative to the glossy red, is all due to differences in saturation
and not to differences in hue. Figure 6.10 shows a portion of the CIE
1931 chromaticity diagram for the red region with the locations of the
chromaticity coordinates from table 6.3 for the three red samples (mea-
sured SCE). Point C is the illuminant point (0% purity) . The lines of
constant purity for 60%, 80%, and 100% are indicated on the di a-
gram and illustrate the decrease that takes place as the surface gloss is
decreased. Clearly demonstrated is the fact that the dominant wave-
length does not change, as illustrated by the straight-line connection of
all points for the red samples with the illuminant point. The dominant
wavelength for all is 614 nm. This effect has been well documented
(Judd and Wyszecki 1975).
At the beginning of this section on surface reflection, it was
stated that highly textured samples exhibiting so-called macro surface
irregularities would not be the major topic of the initial discussion. The
reflectances of the textured gel-coated laminate samples that had been
prepared and measured for the previous example can be of interest,
ref.
spl.
202
Sample Geom.b x
y
Rig SCE 17.16 8.83
Rim SCE 21.27 12.74
Rlt SCE 20.83 12.68
Gig SCE 6.74 6.87
Glm SCE 10.75 10.97
Git SCE 10.72 10.92
Rig SCI 21.55 13.34
Rim SCI 21.83 13.41
Rlt SCI 20.82 12.78
Gig SCI 11.10 11.34
Glm SCI 11.35 11.58
Git SCI 10.82 11.03
Rlt SCE
Rlt SCI
Git SCE
Git SCI
' R ; red; G ; gray; g ; glossy; m ; matte; t ; textured
bSCE; gloss excluded; SCI ; gloss included
Table 6.3
Red and gray gel-coated plastic laminates
measured with integrating sphere geome-
try (CIE coordinates; illuminant C; 1931
standard observer; color difference equa-
tion, eq. 2.4).
Figure 6.10
CIE 1931 chromaticity diagram of the
three red plastic, gel-coated laminates of
different surface structure, measured with
the specular reflectance excluded (SCE).
When surface characteristics alone con-
tribute to the observed color differences,
the only change in chromaticity is in the
purity (chroma) of the color. The dominant
wavelength (hue) remains unchanged.
Chapter 6
t.E*b
z x y t.L * t.a* t.b* t.C (Std. = GSpl)
1.46 0.6249 0.3216
5.77 0.5345 0.3203 +6.72 -8.33 -16.826 19.00 20.1 6
6.22 0.5242 0.3191 +6.63 -10.00 -18.943 21.42 22.42
8.87 0.2997 0.3056
13.79 0.3027 0.3088 8.025 -0.301 0.305 0.43 8.04
13.66 0.3035 0.3094 7.935 -0.168 -0.125 0.21 7.94
6.86 0.5161 0.3196
6.67 0.5208 0.3199 0.10 2.74 0.942 2 .90 2.92
6.52 0.5188 0.3186 -0.84 2.07 -0.075 2.07 2.07
14.18 0.3032 0.3096
14.53 0.3030 0.3091 0.390 -0 096 -0. 173 0 .20 0.44
13.81 0.3034 0.3092 -0.52 -0.188 -0.024 0 .19 0.55
+0.15 -0.71 -0.991 1.22 1.23
+0.19 0.064 0.056 0 .085 0.21
however. The textured surfaces resembled rough leather designed for
possible use as a nonskid surface. The reflectance measurements and CIE
data presented at the bottom of table 6.3 illustrate that there is relatively
little difference on the textured samples, T, between the measurements
made with the specular reflection included or excluded. Generally, both
CIE chromaticity diagram
red plastic gel coats (measured SCE)
0.35
::.._
0.30
D matte
0.25
0 textured
glossy
0.20
0.30 0.35 0.40 0.45 0.50 0.55 0.60 0.65
x
Special Topics 203
of these measurements are a little lower than the total reflectances of the
other two surfaces-the gloss, G, and matte, M-probably due to light
trapping in the irregularities of the surface structure (Ogilvy 1991).
The assumption made for the samples exhibiting micro surface
scattering-that the surface reflectances exhibited rotational symmetry-
that is, that they did not change as the sample was rotated through 360 in
the sample plane-is not a valid assumption in the case of macro textured
samples-for example, textiles or surfaces formed by brush marks (Judd
and Wyszecki 1975). Such mechanically formed surfaces can exhibit char-
acteristic rotational surface reflectance differences. Thus, at constant illu-
mination and viewing angles, the surface reflectance may change as the
sample is rotated in its plane through 360. The most sensitive measure-
ments of such changes can be made with goniospectrophotometry; the
measurements are made at the specular angle, such as -45/45 incident
and viewing angles, and the samples are rotated in the plane through 360.
Although the interpretation of such goniospectrophotometric
measurements may require a high degree of sophisticated knowledge
regarding such materials as textiles, which are not cast against a mold,
characterization on mechanically created surfaces can be informative. For
example, textured papers may exhibit rotational surface reflection that is
characteristic of the calender used in their manufacture.
There are many references in the modern literature relating to
the characterization of the reflectances of surfaces, a problem indepen-
dent of the study of color, and no attempt is made here to survey them.
Numerous mechanical and optical methods have been devised to provide
profiles of the structure of surfaces (Lettieri et al. 1991).
Types of gloss
Attempts to characterize the surface appearance of objects and materials
have been carried on for many years, concurrently with attempts to char-
acterize and describe their color. Hunter ( 19 3 8) and Hunter and Judd
(1939)-see also Judd and Wyszecki (1975)-list five different types
of gloss: specular, sheen, contrast, distinctness of image, and absence
of bloom. All are attributable to aspects of the reflected light basically
caused by the effect of the refractive index and the surface characteristics
on the surface reflectance.
Special instruments have been designed for the measurement of
these various surface reflection factors (Judd and Wyszecki 1975)-as
opposed to absolute surface reflectances measured relative to a perfect mir-
ror. These special instruments are called glossmeters and are, in effect,
abridged goniophotometers. They are designed to measure the relative
intensity of the specularly reflected light at fixed angles of illumination and
viewing. Each instrument is supplied with reference standards calibrated as
unity for the geometrical conditions used. Thus, the measurements are not
absolute and so are not relatable to the effect of gloss on observed color.
For measuring the specularly reflected light, 20 gloss instruments are used
for measuring high-gloss samples; 45 or 60 gloss-measuring instruments
are used for intermediate, semigloss materials; and 75 or 85 glossmeters
are used for near-matte, diffusely reflecting samples to measure their so-
called low-angle sheen (ASTM Standard D 523).
204 Chapter 6
The appearance aspects of the five different types of gloss
have been defined in terms of glossmeter measurements. Specular gloss
is described as it has been used here: reflectance at the angle equal and
opposite to the incident angle (-i
0
/+v
0
, where i = v). Sheen is described as
the shiny reflectance on matte surfaces when observed near the grazing
angle, such as 85 from the perpendicular (-85/85) . Contrast gloss
relates the reflectance at the specular angle to the reflectance well off the
specular angle (such as -60 incident angle and +60 viewing angle ); the
ratio relative to a measurement at 0 (-60/0, that is, -60/60/-60/0)
is then measured. Distinctness-of-image gloss is measured by viewing the
glossy sample at another angle, a few minutes of arc away from the spec-
ular angle, essentially determining the slope of the change in reflectance
at an angle close to the specular. Bloom is a term generally associated
with a haziness that develops on the surface of high-gloss paint films or
plastics, sometimes attributable to the exudation of a plasticizer, for
example. Bloom is measured by making one measurement at the specular
angle, such as - 60/60, and a second measurement a few degrees of
arc away from the specular angle, such as -60/63, and is generally
described, for example, as the ratio -60/60/-60/63. In this case, the
measurement at the second angle is farther from the specular angle than
for distinctness-of-image gloss.
The preferences for an ideal or desired set of characteristics of
a glossy surface are highly subjective. In general, in articles of commercial
importance, if a surface of higher refractive index, exhibiting high inten-
sity of specular reflectance, is flawed-that is, it exhibits some degree of
micro or macro irregularity (like an orange peel )-the imperfect surface
has been found to be more objectionable than a more nearl y mirrorlike
image reflector with no flaws-that is, of better distinctness of image.
This is so even though there may be lower intensity in the reflected image
because of a lower refractive index in the latter case (Judd and Wyszecki
1975). For example, at one time in the automotive industry, it was
thought that the magnitude of the intensity of the surface reflection of a
car's finish was paramount. Later, the trend turned toward placing more
importance on the distinctness-of-image gloss (Braun 1991).
Determining the subjective preference between two samples,
one with sharp distinctness-of-image gloss and the other with higher
intensity of reflectance but lower distinctness of reflected image, poses
problems. Attempts at ranking samples with varying degrees of the two
types of gloss have proved difficult. "In any case the result is dependent
on the specific choice of the observer. The situation is similar to the com-
parison of colored surfaces of the same hue but differing at the same
time both in lightness and in saturation. In such cases the samples cannot
be ordered on a linear scale. We need at least two dimensions" (Seve
1993) . Thus, the problem of ranking samples in terms of their disti nct-
ness of image or the intensity of specular reflectance cannot easily be
reduced to one simple factor or measurable characteristic. Bartleson
(1974), in a report to the CIE, pointed out that gloss and appearance
aspects consist not only of the physical aspect but also of physiological
and psychological aspects associated with vision and experience.
Special Topics
Surface changes after exposure
Changes in the surface reflection of materials often occur during aging
or weathering, or after use (see remarks on surface reflection changes
205
in Johnston and Feller [1967] and Johnston-Feller and Osmer [1977]).
The process of change can be so gradual as to be overlooked unless the
exposed areas can be compared to an unexposed section of the original
material. The general trend of these surface changes is to approach a
semimatte surface; on high-gloss materials, the result of exposure is
frequently to lower the gloss; while on matte samples the trend is to
increase the gloss due to burnishing. Actually, it is extremely difficult to
clean most matte, nonglossy surfaces without increasing the gloss. It is
essential to remember the importance of surface changes on perceived
color changes. In general, it can be stated that the magnitude of any
change in the surface must be taken into account before real changes in
pigmentation can be accurately determined.
Fluorescence
Fluorescence is a process in which a material absorbs electromagnetic
flux of particular wavelengths and emits radiation at other, longer wave-
lengths. Phosphorescence and fluorescence are two types of the general
phenomenon of luminescence. They differ in the length of time that
reradiation occurs: phosphorescence continues noticeably after irradiation
ceases; fluorescence does not. There are many types of luminescence based
on the source of the absorbed energy (Grum 1980; Byler 1973). Here we
consider only the fluorescence of organic compounds caused by ultravio-
let, visible, or near-infrared radiation. For example, ultraviolet radiation
may be absorbed and some of the energy emitted in the visible region;
visible radiation may be absorbed and some of it emitted in either the
longer-wavelength visible or the near-infrared region. These are the partic-
ular situations that will be discussed. For those interested in the origin of
such fluorescence in a compound, Voedisch (1973a, 1973b) and de la Rie
(1982) present explanations of the molecular structures responsible.
As we emphasize throughout, the color of mixtures of most
dyes and pigments is the result of subtractive colorant mixture, depen-
dent on the absorption and scattering coefficients of the systems. In the
case of fluorescent colorants, however, the perceived color is the result of
the additive mixture of the emitted fluorescent radiation with the accom-
panying normal reflected color of the colorants and, possibly, the sub-
strate (Voedisch 1973a). An illustration of this phenomenon is given in
figure 6.11, which shows the spectral curves of a white plastic without a
fluorescent blue whitener (fluorescent whitening agent, FWA), curve 1,
and the same plastic with an FWA added, curve 2. Curve 3 shows
fluoresced radiant energy that is additively mixed with the reflected light
of curve 1. The light source used to irradiate the samples was an approx-
imation to CIE illuminant D65 (daylight), which is defined in the ultra-
violet region as well as in the visible. All were measured relative to a
pressed BaS04 standard. The term radiance is used to describe the
206
Figure 6.11
Spectral curve measurements of a white
plastic: curve 1-without a fluorescent
whitening agent (FWA); curve 2-with
the addition of an FWA; curve 3-the
fluoresced radiant energy from the FWA
(after Grum 1980) .
Chapter 6
-v B G
y
0 R
130
2
120
110
100
M'
C:!.
C1J
90
u
c
~
"

....,
80 c
C1J
~
C1J
Cl.
0 70
~
C1J
u
c
60
cu
....,
u
C1J
'+=
~
50
....,
c
C1J
3

C1J
CL
40
30
20
10
400 450 500 550 600 650 700
Wavelength (nm)
fluorescent emission or the reflectance plus the fluorescence, while the
nonfluorescent portion is t ermed simply the reflectance. However, in pop-
ular usage the simple term reflectance is often used instead of radiance.
Color measurement of fluorescent materials
Because the observed reflected color is the result of the emittance at
wavelengths different from a particular wavelength incident on the
material, a special optical arrangement in spectrophotometers or col-
orimeters must be used for measuring the perceived color of materials
that fluoresce. The fluorescent sample must be illuminated by polychro-
matic light (white light), and the reflected light (or transmitted light in
the case of transparent materials ) plus the emitted light analyzed with a
monochromator or filters (as mentioned in chap. 1 ). In the past some
spectrophotometers illuminated the sample with monochromatic illumi-
nation. Illumination of fluorescent materials by monochromatic light
results in color measurements very different from the colors observed
Special Topics 207
visually under normal polychromatic light sources. Incident light at the
wavelength of excitation is registered by the instrument's detector as
reflected light at that particular wavelength and not at the wavelength of
the fluorescent emission. It is, therefore, important that the user know
whether his instrument uses polychromatic or monochromatic illumina-
tion, so that fluorescent samples can be measured in accordance with the
visual appearance.
Another special problem when measuring highly fluorescent
samples is that the light reflected by these materials is often well over
100% relative to that reflected by the normal white diffusing reference
standard. Some measuring instruments, however, may not record over
100% or 120%. In such a case, a neutral-density filter of known trans-
mittance can be inserted into the sample beam to decrease its intensity by
a known amount. Thus, a 20% neutral-density filter decreases the inten-
sity to 20% of the actual radiance so that measurements recorded will be
on scale. Such decreased reflectances can then be increased by the appro-
priate factor to establish the actual values of the radiance.
There is one other facet of instrument design that is impor-
tant in measuring fluorescent colors accurately: the use of an integrating
sphere compared to the use of a bidirectional, geometric illuminating and
viewing arrangement. The latter geometric arrangement is recommended
for fluorescent samples because in the sphere arrangement, the color of
the fluorescent sample may alter the character of the incident light, thus
altering the color of the reference white. The effect decreases as the area
of the sample measured is decreased relative to the area of the sphere
wall (ASTM Standard E 991). Based on this author's experience with a
few good commercial sphere-type color-measuring instruments, the effect
is not one of concern for ordinary usefulness. Giving up the ability to
either include or exclude the specular reflection, as is possible on most
good sphere instruments, is far more important than any small error in
accuracy that might occur. However, the author has no experience with
instruments utilizing small spheres where the sample size illuminated may
be very large relative to the sphere size.
Historical use of fluorescent pigments
A number of organic fluorescent compounds have found their way
into museum objects; some are many centuries old, others are of recent
vintage. One of the oldest organic fluorescent pigments to be found is
natural madder lake, which is based on alizarin ( 1-2-dihydroxyan-
thraquinone) from the root of the perennial plant Rubia tinctorum.
This pigment is a lake of alizarin, generally encountered as the alumina
lake but also containing an amount of purpurin (1,2,4-trihydroxyan-
thraquinone), which fluoresces a bright yellow-red. Feller (1968 ) has
published the curve of the fluorescence of natural madder activated by
energy of 400 nm wavelength, as well as the curves for weaker fluores-
cence activated by 345 nm. Feller's curve of the fluorescence activated by
400 nm incident light is redrawn here as figure 6.12. As can be seen, the
peak of the fluoresced radiance is in the 580 nm yellow region. The lake
made from pure alizarin, manufactured synthetically since the latter part
of the nineteenth century, is not fluorescent since it contains no purpurin.
208
Figure 6.12
Spectral relative fluorescent radiance of
natural madder activated by 400 nm
radiation (after Feller 1968).
Chapter 6
90
80
70
<!)
u
60 i:::
~
"D
::Ii
....,
i:::
<!)
u
"'
50 <!)
0
:::J
q:::
<!)
>
+:;
m
40 Q)
er:
30
20
10
500 540 580
fluorescence of natural madder
activated by 400 nm incident light
620 660 700
Wavelength (nm)
740
Another traditional organic fluorescent pigment to be found
in artworks is Indian Yellow, made from the urine of cows fed mango
leaves. The principal component is the yellow, crystalline magnesium salt
of euxanthic acid. Baer and coworkers (1986) state that it was used in
Indian paintings beginning in the late sixteenth century, and it has been
identified in European paintings beginning in the late eighteenth century.
Since the early twentieth century, however, manufacture has been prohib-
ited. Feller (1968) and Baer and coworkers (1986) have measured the
fluorescence radiance curve of Indian Yellow excited by incident radiation
of 436 nm. The fluorescence peaked at 535 nm in the yellowish green
region. Figure 6.13 is a redrawing of Feller's fluoresced radiance curve.
Figure 6.14 is a redrawing of Feller's total radiance curve as measured on
a Model D-1 Color-Eye abridged spectrophotometer. In this instrument,
the sample is illuminated with an incandescent lamp, which emits little
ultraviolet radiation to be absorbed and reradiated as fluorescence. The
Figure 6.13
Spectral relative fluorescent radiance of
Indian Yellow activated by 436 nm radia-
tion (after Feller 1968; Baer et al. 1986).
Special Topics 209
y 0 14---- R----
90
80
fluorescence
Cl>
induced by 436 nm u
70
c
<U
'6
[<j
""O
Cl>
:t:j
60
E
Cl>
'+-
0
>- ....,
v;
50
c
2
c
Cl>
>
+='
<U
40
QJ
er:
30
20
10
400 450 500 550 600 650 700
Wavelength (nm)
light reflected is then analyzed by means of narrow band-pass filters. This
optical arrangement is the correct one for including the fluorescence in
the measurement. Figure 6.15 is a redrawing of Feller's curve of the
relative fluoresced radiance at 535 nm as a function of various incident
activation wavelengths from 312 nm to 436 nm. It can be seen that wave-
lengths in this ultraviolet region are not necessary to activate the fluores-
cence. Strong fluorescence is activated by light in the visible region at
436 nm. Under normal illumination conditions, this fluorescent light is
additively mixed with ambient light so the observer is unaware of its
occurrence. An ultraviolet light source is normally used to detect the
presence of Indian Yellow. Baer and coworkers (1986) show colored pho-
tographs of an Indian miniature painting of the early eighteenth century,
one illuminated with visible light and another with near-ultraviolet light.
The presence of Indian Yellow is immediately suspected because of the
greenish color of the fluorescence that is visible in the photograph taken
under ultraviolet illumination.
210
Figure 6.14
Total spectral radiance of Indian Yellow as
measured on a Model D-1 Color-Eye with
incandescent illumination on the sample.
Chapter 6
y 0 ----R----
90
80
'\
70
I \
I \
\
I
\
I
\
cu
60
\ u
I c:::
\
.l'i I
u
\
cu
I
q::
\
I

50
I
\
.8 I
\
.....
I
\
c:::
\ cu
I

\ cu
I CL
40
\
I
\
I
I
\
I
\
30
I
\
I
\
\
I
\
I
\
20
I
\
I
\
I
\
I
\
I
\
10
I
\
'
400 450 500 550 600 650 700
Wavelength (nm)
Development of modern fluorescent, high-visibility pigments
During World War II, the combating countries all made use of fabric sig-
nals coated with fluorescent dyes because of their high visibility (Switzer
and Switzer 1950). Nearly all of the dyes were extremely fugitive, how-
ever. Though pigments had been made by incorporating fluorescent dyes
into a resin matrix and grinding the solid mixture into a powder, it was
not until Switzer developed the technique of incorporating the fluorescent
dyes into urea and melamine resins that their commercial use as pigments
began. The patent was issued in 1950 (Switzer and Switzer 1950).
The early pigments had only limited light stability because of
the fugitive character of the dyes used; they faded after about one month
of exterior exposure. Nevertheless, they were quickly utilized by the
advertising industry because of their high visibility. Their value for
other uses, particularly for safety signs and directions, was soon realized
because of their attention-attracting ability. The value of such applications
hastened development by the manufacturers of more light-stable matrices.
Figure 6.15
Relative fluorescent radiance of Indian
Yellow at 535 nm as a function of various
incident activation wavelengths from
312 nm to 436 nm.
Special Topics 211
100
90
80
E
c
l{)
M
70
l{)
-:;;
"
QJ
~
E
QJ
60
QJ
u
c
~
"
.....,
50
c
QJ
~
QJ
0
~
;:;::::
QJ
40 >
:p
<1l
Qj
0::
30
D
,,
I \
I
\
I
20
I
\
10
300 350 400 450 500 550 600
Wavelength of incident radiation (nm)
The improvement in stability was obtained by incorporating ultraviolet
absorbers into the resin. One of the first applications was for signs and
markings on military aircraft and then on civilian aircraft, using in partic-
ular the familiar aircraft orange. Voedisch (1973b) cites and describes mil-
itary specifications for this material.
Figures 6.16-6.20 illustrate spectral measurements of paper
sheets coated with daylight fluorescent pigmented coatings purchased
from an artists' supply store. The colors were fluorescent green, yellow-
green, orange, pink, and red. Three measurements for each coating are
illustrated in the figures. In the two measurements made with the Color-
Eye 1500 abridged spectrophotometer, the sample is illuminated with
polychromatic illumination, and the total observed light-that is, the
reflected plus the fluorescent emitted light {the radiance)-is recorded.
The illuminant used is a pulsed xenon lamp filtered to approximate CIE
illuminant D65. The instrument is also equipped with an ultraviolet filter
212
Figure 6.16
Spectral curves of a fluorescent green
coating on paper measured in three differ-
ent ways. The measurement made with the
Trilac spectrophotometer, which illuminates
the sample with monochromatic light, des-
ignated 8I D, is obviously not correct. The
measurements made with the Color-Eye
1500, which illuminates the sample with a
polychromatic xenon flash, designated
D/ 8, indicate the magnitude of the fluo-
rescence. Note that removing most of the
ultraviolet wavelengths by means of a filter
decreases the fluorescent radiance by a rel-
atively small amount. The CIE tristimulus
values and chromaticity coordinates are
presented in table 6.4.
Chapter 6
200
180
160
Q) 140
u
c
. ~
"'Cl
~
....,
~ 120
~
Q)
a..
} 100
g
Q)
u
c
~
Q)
;:;:::
!'::'
....,
c
Q)
~
Q)
CL
80
60
40
20
400
G
y 0 14-----R----
fluorescent green
--- Trilac (8/ D)
- - - Color-Eye 1500 (D/8') UV-in
-- Color-Eye 1500 (D/8' ) UV-out
~
I \
1/\\
,. \\
,, .,,

I/ \ \
,. Q
,, ,,
. \
I/ \\
'i ~
I. ~
,, ''*
. ~
6 ~ ~
I
I
I
I
I
I
~
~
.
I
I
...... p
0...-0- O' /
450 500 550
Wavelength (nm)
600 650 700
(UF-4) to remove all of the ultraviolet portion of the source spectrum
below 387 nm (it transmits only 60% at 400 nm). The graphs show the
reflectances measured both ways, designated as "UV-in" for the full D65
daylight illumination, and "UV-out" when the UV filter was in place
between the light source and the sample. The third measurement on the
graphs is the reflectance curves obtained from the Trilac spectrophoto-
meter. This illustrates the situation in which the sample is illuminated
with monochromatic (si ngle wavelength} illumination rather than with
the polychromatic illumination required to measure both the fluorescence
and the normal diffuse reflectance, as used in the Color-Eye 1500.
The CIE notations for the curves illustrated in figures
6.16-6.20 were calculated for illuminant C and the 2 standard observer
for each of the three measurements made on each coating. They are
presented in table 6.4. The luminous reflectance of the yellow-green is
greater than 100%; for others, the reflectance at some wavelengths is
also greater than 100% relative to the usual diffuse white standard.
Table 6.4
CIE notation for fluorescent samples illus-
trating measurement differences (illumi-
nant C, 1931 standard observer).
Special Topics 213
With the monochromatic illumination used in the Trilac, as
well as in many older or conventional spectrophotometers and colori-
meters, the emitted fluorescent energy is registered by the instrument
detector as being reflected at the excitation wavelength (note in figs.
6.16-6.20 the region of maximum absorbance in each case). Reflectance
measurements such as these made on the Trilac bear no resemblance to
what one would see. The optics of this type of instrument, therefore, are
inappropriate for measurements made to describe what we see on materi-
als that exhibit fluorescence in the visible region of the spectrum. (See the
section on instrumentation for color measurement in chap. 9. ) Most mod-
ern instruments designed for color measurement utilize the polychromatic
illumination that is correct for fluorescent samples. Some expensive instru-
ments are built with reversible optics, which allow the user to use either
polychromatic or monochromatic illumination for spectral measurements.
There is something to be learned by comparing the curves
made with monochromatic illumination (the Trilac measurements in figs.
6.16-6.20) to those made using the Color-Eye Model 1500 with poly-
chromatic illumination. The wavelengths of reflectance on the Trilac
curves (monochromatic illumination), which are high, relative to those
made with the Color-Eye (polychromatic illumination), reveal the wave-
lengths of importance in the excitation of the fluorescence. For all of
these high-visibility pigments, visible wavelengths excite the fluorescence.
The presence of ultraviolet wavelengths is not necessary for exciting the
fluorescence. Thus, excluding the ultraviolet region from the illuminant
does not decrease the brilliance of the emitted fluorescence by a large
amount. In addition, the ultraviolet-filtering compounds in the pigment
resin matri x absorb most of the ultraviolet. Hence, a museum object
containing these daylight fluorescent colorants is not likely to be altered
in appearance to a great extent by the use of ultraviolet filters over the
light source when it is displayed. Thus, another technique for protecting
the fluorescent dyes from fading is the application of a clear varnish con-
taining ultraviolet absorbers over the surface.
Sample Optics x
y
z x y
Green UV-in 53.12 92.57 17.25 0.3260 0.5681
UV-out 50.78 87.09 16.25 0.3295 0.5650
Tri lac 51.78 66.31 67.1 2 0.2796 0.3580
Yellow- UV-in 86.52 116.40 23.04 0.3829 0.5151
green UV-out 82.55 109.64 22.1 2 0.3852 0.5116
Tri lac 85.06 89.44 79.74 0.3346 0.3518
Orange UV-in 107.23 89.63 9.35 0.5200 0.4347
UV-out 98.95 82.69 9.23 0.5184 0.4332
Tri lac 77.54 74.55 61 .27 0.3634 0.3494
Pink UV-in 78.18 42.02 53.97 0.4488 0.2413
UV-out 71 .97 38.75 53.60 0.4380 0.2358
Tri lac 65.79 54.06 72.1 9 0.3426 0.2815
Red UV-in 81 .14 48.27 8.47 0.5885 0.3501
UV-out 75.51 44.98 8.53 0.5852 0.3486
Tri lac 62.63 54.94 45.98 0.3830 0.3359
' UV-in and UV-out measurements made on the Color-Eye 1500 abridged spectrophotometer.
214
Figure 6.17
Spect ral curves of a fluorescent yellow-
green coating on paper measured in the
three ways as described for figure 6.16.
Chapter 6
200
180
160
Q) 140
u
c
~
"

....,
c
120 Q)
~
Q)
0..
0
I
100
0
;t::
Q)
u
c
~
u
Q)
80
""

....,
c
Q)
u
Q;
Cl..
60
40
20
400
,.
,!
,,
,.
,!
,,
,.
Ji
t
'O...._ /
o ~
y 0 ---- R----
fluorescent yellow-green
--- Tri lac (8' /D)
- - - Color- Eye 1500 (D/8' ) UV-in
-- Color-Eye 1500 (D/8' ) UV-out
450 500 550 600 650 700
Wavelength (nm)
These comments aren't meant to be misleading: it must be
remembered that any fluorescent material, such as those made with nat-
ural pigments or dyes, should be protected from prolonged exposure to
ultraviolet wavelengths.
Uses of fluorescent, high- visibility pigments by artists
As the use of fluorescent colorants expanded, artists soon became fasci -
nated by these "new" pigments with which they could create innovative
and unusual optical effects. By the 1960s artists were using them despite
their limited light stability, which was well publicized by scientists as
well as by the manufacturers at that time.
The adoption of these pigments by artists necessitated their
learning to mix colorants a little differently from mixing absorbing,
nonfluorescent pigments. For example, if a fluorescent pigment is mixed
with a nonfluorescent pigment that absorbs in the spectral region that
Figure 6.18
Spectral curves of a fluorescent orange
coating on paper measured in the three
ways as described for figure 6.16.
Special Topics
200
180
160
<!) 140
u
c
ro
'6
~
...,
c
120 <!)
~
<!)
0..
0
<ii
100
g
<!)
u
c
_:g
u
<!)
80
'+=
~
...,
c
<!)
~
<!)
CL
60
40
20
' I
0 D
-o- -0-
400 450
215
y 0 -----R----
/
,Cf'
(y
I
I
I
I
I
~
.... ,
I .. ,
19 '\ \
, . ~ \
,! \ \
I I . \
,. \ ~
I! q \\
,, \ \
,. . '
,! C\. ''
,, '--"'
~ ~ '
fluorescent orange
--- Trilac (8/D)
- - - Color-Eye 1500 (D/8) UV-in
- - Color-Eye 1500 (D/8) UV-out
500 550 600 650
Wavelength (nm)
700
excites the fluorescence or in the spectral region where the fluorescence is
emitted, the fluorescence is decreased or quenched.
In an article describing his use of fluorescent pigments in his
paintings, the artist Herbert Aach (1970) wrote, "I have been painting
with fluorescent pigments for over five years. Only now do I begin to real-
ize that it took me over three years to relearn color behavior with such
pigments." Because the hiding power of these fluorescent pigments is low
because of their transparency (they are made with dyes incorporated into
a transparent matrix), the reflectance of the substrate, whether neutral or
chromatic, as well as the absorption and scattering characteristics of any
colorant with which they may be mixed, is of major importance. Aach
had to adapt to these factors in relearning colorant behavior.
Because the fluorescent pigments available at that time con-
tained ultraviolet absorbers in the resin matrix, they are not usually acti-
vated by black-light illumination (an ultraviolet source filtered to remove
216
Figure 6.19
Spectral curves of a fluorescent pink coat-
ing on paper measured as described for
figure 6.16.
Chapter 6
200
180
160
"'
140
u
c
~
"
r:
.....
c
120
"'
"'
a_
0
]
100
g
"'
u
c
.s
u
"'
80
""

.....
c
"'
"'
CL
60
40
20
400
y 0 ----- R----
fluorescent pink
--- Trilac (8/ D)
- - - Color-Eye 1500 (D/8) UV-in
-- Color-Eye 1500 (D/8) UV-out
450 500 550
Wavelength (nm)
1"
I \
,fb \
. \ \
,, . '
,. \ \
,1 Q \
ti \ ~
J; \ ~ ',
~ ..
r \ ,
I 0..
I
I
I
I
I
I
I
600 650 700
visible wavelengths) alone. Aach circumvented this limitation by making
up his own fluorescent paints as he needed them, using an extremely
high-pigment-volume concentration. Commercial paints cannot be for-
mulated this way because of extreme settling problems in the container.
Use of this very highly concentrated pigmentation, however, retained
sufficient ultraviolet response to be effective aestheticall y; it also resulted
in improved light stability.
In a subsequent article, Aach (1972) states, "There are by
now a goodly number of us [creative artists] who are working in fluores-
cent color. Frank Stella, Ralph Humphrey, Al Loving, Frank Bowling
come instantly to mind."
The author has a print made by Aach containing fluorescent
orange pigment in addition to nonfl uorescent pigments. The print,
framed under glass, hung on her office wall at Carnegie Mellon Research
Institute for seventeen years . The wall on which it hung was illuminated
Figure 6.20
Spectral curves of a fluorescent red
coating on paper measured as described
for figure 6.16.
Special Topics 217
y 0 ----R----
180
160
cu
140
u
c
~
"O
~
~
c
120 cu
u
QJ
c..
0
<ii
100
g
cu
u
c
l'l
u
cu
80 :;::
~
~
c
cu
~
cu
a_
60
40
20
400
fluorescent red
--- Trilac (8/D)
- - - Color-Eye 1500 (D/8) UV-in
-- Color-Eye 1500 (D/8) UV-out
'"
I \
I 1'\ \
.. \'
Ir . -
6 \\
,. b \
I ' ' I. \ ~
I/ . \
,. b \
1! \-
1/ 'o..'
,. .
1!
,,
,.
1!
I/
,.
Tj
11
I/
l'
l
I
' d a.. -o.._... /
'o- -o- a v- -o
450 500 550 600
Wavelength (nm)
650 700
by diffuse daylight from a nearby window and by ceiling fluorescent
luminaires. When the print was recently removed from its frame, the
fluorescent paint showed no obvious signs of fading. It was still activated
by ultraviolet light alone, as illustrated in color plate 5.
Whether other artists who used fluorescent colorants for-
mulated their paints or inks as well as Aach is not known. Hence, any
objects containing fluorescent colorants should be considered to be
sensitive to prolonged expos ure to light, particularly ultraviolet.
Fluorescent whitening agents (FWAs)
No mention has been made above of blue fluorescent pigments being used
for high-visibility applications. Because our eyes have far lower sensitivity
to wavelengths of light near the ends of the spectrum-the long-wave-
length deep reds and the short-wavelength blues and violets-colorants
that fluoresce in these regions are not among the fluorescent dyes and
218 Chapter 6
pigments popularly called high-visibility colorants. In the case of the
violet-blue fluorescent dyes, their particular and very widespread applica-
tion is as FWAs to enhance the whiteness appearance of yellowish whites.
Many organic materials become more yellow with age. In
other cases, materials such as paper and resins that are intended to be
white or colorless are often yellowish. The reader, if too young to remem-
ber, will learn from history that one of the most widespread applications
for ultramarine blue pigment was as a bluing agent used in laundering
and in neutralizing the yellowness in other products. The employment of
such a blue pigment, however, lowered the resulting reflectance; it neutral-
ized the yellowness by absorption in the yellow-orange region, resulting in
graying rather than in a true whitening effect. Today violet-blue fluores-
cent dyes produce a neutralizing effect by additive color mixture, actually
brightening the color while simultaneously neutralizing the yellowness.
Figure 6.11 illustrates the effect: the addition of a blue whitener dye has a
significant impact on the color of the white plastic. It can be seen that the
total radiance curve of the white plastic plus the fluorescent blue whitener
is far higher than the reflectance of the white plastic alone.
Most modern laundry detergents contain these fluorescent
brighteners; many white paper products or white plastic materials also
contain them. When measuring the discoloration or whiteness of paper,
one must take this possibility into consideration. One has only to observe
white laundered textiles or white papers under a black-light ultraviolet
source to confirm the brilliance contributed by these products. This use of
blue-violet fluorescent dyes or pigments is a special and somewhat sepa-
rate application, certainly very common in recent years. The exciting radi-
ation (for the blue-white fluorescence) is mostly in the ultraviolet region.
Hence, eliminating this region by absorption using ultraviolet filters or by
mixture with ultraviolet-absorbing pigments such as rutile Ti02 destroys
the fluorescence. One does not find FWAs used in paints or plastics pig-
mented with Ti02, either rutile or anatase, which absorbs to a greater or
lesser extent some of the violet and most of the ultraviolet region.
The whole subject of whiteness and FWAs has been gi ven an
excellent review by Grum (1980). He gives extensive references to other
workers who have published on the subject. The reader is referred to the
section on single-number color scales in chapter 4 for a brief description
of whiteness indices.
Weak fluorescence of resins
No mention has been made up to this point of the fluorescence that
occurs as many natural vehicles and varnishes age. Linseed oil, damar,
and mastic are examples that not only yellow but also develop fluores-
cence as they age. Examination of paintings under ultraviolet light is
routinely used to detect retouches done at a time later than the original
painting, for example. For an informative study of this phenomenon, the
articles by de la Rie (1982) should be consulted.
In relation to the degree of fluorescence exhibited by the day-
light, high-visibility colorants or the FWAs, the amount of fluorescence
of old varnishes and oi ls is generally extremely small and visible only
in a darkened room under black light. Under normal illumination the
Special Topics
observer would not be aware of its existence. It is doubtful that it is
sufficiently strong to affect the spectral measurements that might be
made on the colorants of old oil or varnished paintings, regardless of
219
the type of illumination-monochromatic or polychromatic-used in the
color-measuring instrument. For safety's sake with paintings that are old,
however, precaution should be taken when they are being measured with
an instrument that illuminates the sample with polychromatic light con-
taining ultraviolet wavelengths, such as a xenon source. The instrument
should be operated with a filter over the light source to eliminate the
wavelengths that activate ultraviolet fluorescence.
It was said at the beginning of this section that the phospho-
rescence of inorganic materials would not be discussed. However, many
objects in museum collections contain these beautiful materials, and it
is possible that many more objects of interest containing various phos-
phors will be preserved. The techniques of spectral measurement
described may be of value in establishing identification procedures.
Spectral character as well as the decay time of emission may provide
identifying parameters.
Microvoids and Vesiculated Beads
Another category of "pigment" deserving special mention may not con-
tain any pigment at all but consists chiefly of air-filled voids suspended in
a polymer matrix. It has been mentioned previously that entrapped air
can act as a good light scatterer. Figure 4. 7 illustrates such a situation.
In this example, the development of voids was unwanted. However, air
spaces in a polymer matrix can be created intentionally, providing an
inexpensive way to introduce desirable scattering in a suspending
medium and thus minimizing the need for white pigment.
Many of the white substances encountered in nature do not
contain white pigment: snow, clouds, white flowers, foam, and suds
(whether on beer or soap solutions) contain no pigment-they contain
only air mixed with finely divided particles. Air may not be the only
cause of the scattering. Similar scattering phenomena occur in milk and
cooked egg white; they contain no white pigment. In his article "Hiding
without Pigments," Burrell (1971) gives an excellent explanation of these
and similar phenomena and gives a short history of some of the applica-
tions of the scattering phenomena created by refractive index differences
between particles.
The basic principle stems from the Fresnel equation (eq. 4.3),
which relates the mirrorlike reflection, p, arising at the interface between
two materials of different refractive index, n
1
and n
2
, where n
1
is the lower
refractive index and n2 the higher. For light perpendicular to the interface,
(n2 - ni)2
p = ----
(n2 + ni)2
(6.1)
where p is the decimal fraction of the incident light that is reflected at
the interface at an angle equal and opposite to the incident angle.
220 Chapter 6
For light not perpendicular to the interface, the angle of inci-
dence in the less dense medium and the angle of refraction in the more
dense medium relate to the interfacial reflection in the following way:
(6.2)
where i is the angle of incidence, and r is the angle of refraction.
Here the refractive indices are given by the basic Snell's law
relationship, n2/n1 =sin i/sin r. All the above equations assume that the
incident light is not polarized. For polarized light, see the equations in
Judd and Wyszecki (1975).
Ordinary pigments hide (scatter light ) largely because of
their high refractive index relative to that of the vehicle. Microvoids
hide (scatter light) because of the extremely low refractive index of air
in the void, relative to that of the vehicle matrix in which they are sus-
pended. If one looks at the ratios of refractive indices of scattering pig-
ments and air-filled microvoids, it can be seen that entrapped air pockets
can indeed be very efficient sea tterers. By use of 2 .5 as an average
refractive index for scattering pigments and 1.5 as an average refractive
index of organic vehicles, the ratio of the pigment to vehicle (higher to
lower refractive index) is 1. 7. For the same vehicle with air-filled voids,
the ratio of the higher refractive index-the vehicle in this case-to the
lower refractive index for air (1.0) is 1.5. It is apparent that air-filled
voids are potentially very good scatterers of light. Because the refractive
index ratios are so similar, the desired size of pigment particles as well
as of voids to achieve maximum scattering is about the same-less than
1 m in diameter. Recall the universal scattering curve, based on Mie
theory, in figure 4.29.
Burrell ( 1971) describes three methods for making microvoid
coatings, only the last of which he considers to be important: formation
of a gel structure containing a nonsolvent for the vehicle in the liquid
phase. He presents a typical paint formulation and outlines the basic
principles for selecting a polymer and the two solvents, a volatile non-
solvent and a less volatile polymer solvent. Many such combinations
have been reported; the patent issued to Seiner (1970) and assigned to
PPG Industries contains hundreds of formulations.
Microvoids formed in situ as a polymer film dries may not
be (and often they are not) the sole scattering source; instead, they are
used in combination with Ti0
2
In this case, the total scattering may be
greater than the sum of that from the two individual scatterers, possibly
because of the occurrence of air-Ti0
2
interfaces.
An excellent review of the preparation, properties, and uses
of microvoids in coatings and in other materials such as paper is given by
Rosenthal and McBane (1973). If the comprehensive theory of the optics
of microvoid "pigmented" materials is of interest, the articles by Kerker
and Cooke (1976); by Kerker, Cooke, and Ross (1975); and by Allen
(1973) may be consulted.
Special Topics 221
It is doubtful that this technology has been utilized in the for-
mulation of artists' materials. However, it certainly may be encountered
in commercial products, and artists and artisans who use commercial
paints may, albeit inadvertently, be incorporating microvoids in their
coatings. The popularity of using this technique to improve hiding
(increase scattering) is a strong economic one: air spaces are less expen-
sive than pigments.
The presence of air spaces inside materials may occur unin-
tentionally, for example, due to incomplete wetting of pigments; after the
thorough curing of a polymer matrix; or following exposure to deleteri-
ous conditions. Michalski (1990) describes the action of solvents, used to
clean the surface of a painting (to remove old varnish and so on), that
results in the formation of microvoids due to solvent diffusion into the
lower layers and subsequent evaporation. He refers to the work of
Erhardt and Tsang (1990): "Erhardt and Tsang have measured substan-
tial blanching due solely to solvent application, not abrasion. Whether
the explanation be micro-voids, surface disruption, granular redeposition
or all three remains to be studied."
The monetary value to be realized from using microvoids as
pigments led to the development of preformed microvoids in the form of
hollow plastic spheres containing air; they are called vesiculated beads
(Hislop and McGinley 1978). The adjective "vesiculated" is derived from
the noun "vesicle," used in the biological sciences to refer to a "fluid-
filled pouch, such as a cyst, vacuole, or cell" and in geology to refer to a
"cavity in a mineral or rock." Such preformed microscopic beads may be
hollow or pigmented. The pigmentation may be incorporated in the vehi-
cle forming the beads, with air in the middle, or it may be encapsulated
inside the beads, or both.
One other special product utilizing bead technology is the use
of solid plastic beads, either pigmented or clear, to reduce burnishing-
that is, the increase in surface gloss of nonglossy (matte) materials. A
brief review is presented by Hislop and McGinley (1978).
Although the products and techniques described here may
not be encountered in antiquities, they certainly may be found in late-
twentieth-century objects and materials.
Extenders, Fillers, Inerts
In the preceding section, solid plastic beads were mentioned, as well as
hollow, vesiculated beads. Solid beads can be considered a special type of
extender pigment. The Federation of Societies for Coatings Technology's
Coatings Encyclopedic Dictionary (FSCT 1995) defines "extender pig-
ments" as "a specific group of achromatic pigments of low refractive
index (between 1.45 and 1.70) incorporated into a vehicle system whose
refractive index is in a range of 1.5 to 1.6." Consequently, extenders
do not contribute significantly to the hiding power of paint. They are
used in paint to reduce cost, achieve durability, alter appearance (e.g.,
222
Table 6.5
Common extender pigments.
Chapter 6
decrease gloss), control rheology, and influence other desirable proper-
ties. If used at sufficiently high concentration, an extender may con-
tribute to dry hiding and increase reflectance.
The Coatings Encyclopedic Dictionary states that the word
filler is synonymous with extender. It is in this sense that the word filler
is used here. Additionally, the term inert pigment is defined as being fre-
quently used for extender. Based on these definitions of the terms, the
most widely used extender pigments are listed in table 6.5; the list,
admittedly, is not exhaustive. Patton (1973) presents extensive discus-
sions of the properties and uses of these and a number of other less com-
monly encountered inerts.
In addition to the extender pigments listed in table 6.5, light
aluminum hydrate (Pigment White 24) and gloss white (75% BaS0
4
and
25 % aluminum hydrate, Pigment White 23) can also be encountered in
pigment mixtures. According to Patton (1973), Pigment Whites 23 and
24 are not prepared as independent, dry pigment powders but instead are
used primarily as pigment substrates (often prepared in situ) for the
manufacture of lake pigments. Alizarin is an example of a pigment that
has been laked on aluminum hydrate.
In his book Artists' Pigments: Lightfastness Tests and
Ratings, Levison (1976), founder in 1933 of the Permanent Pigments
Company, discusses the major reasons for the incorporation of extender
pigments in artists' oil paints made with organic pigments. Because the
book was privately printed and is now out of print, the author will quote
many of Levison's remarks verbatim:
Organic pigments differ considerably in physical properties
from the inorganic pigments to which the artist is accus-
tomed. Most inorganic or mineral type pigments are fairly
dense and require but a moderate proportion of oil to make a
durable and proper working artists' paint. Organic pigments,
on the other hand, are light, fluffy and very finel y divided so
Name Chemical type Colour Index Refracti ve index
Barytes natural BaS0
4
PW 22 1.637
Blanc fixe synthetic BaSO
4
PW 21 1.637
Chalk (whiting) calcium carbonate PW 18 1.48-1.65
Dolomite calcium carbonate and PW 18 1.50- 1.68
magnesium carbonate
Precipitated chalk synthetic calcium carbonate PW 18 1.53-1.68
Silica (amorphous) silicon dioxide PW 27 1.54
Silica (diatomaceous) hydrous silica PW 27 1.48
Silica (synthetic) (Si0
2
)x - (H
2
0)y PW 27 1.46
Silica (pyrogenic) silicon dioxide 1.45
Gypsum hydrated calcium sulfate PW 25 1.52-1.53
Talc magnesium silicate PW 26 1.54- 1.59
Kaolin clay hydrated aluminum silicate PW 19 1.56
Mica aluminum potassium si licate PW 20 1.58
Fuller's earth hydrated aluminum 1.50- 1.55
magnesium silicate
' PW = Pigment White
Special Topics 223
that it requires considerable oil to "wet" all the pigment par-
ticles and make a workable paint.
With the inorganic pigments the percent oil in the total
paint will vary from less than 20% in the whites to 24% in
the cadmiums, and up to close to 40% for other pigments.
With a concentrated organic pigment (called a " toner") the
percent oil to make a workable paint can be from 60% to
over 70%. Such an excess of oil produces poor drying, a soft
film, possible wrinkling and excessive yellowing. With acrylic
emulsion paints there is no problem since the pigment is not
critical to the properties of the binding vehicle.
A highly durable [oil] paint, however, can be formulated
with organic pigments by providing a backbone of dense inor-
ganic "inert" pigment. An inert pigment is one that has little
or no coloring power. The ancient and Renaissance artists
made their organic colors (obtained from natural sources)
by absorbing the color on clays or chalks, which are inerts.
Today's oil paints which contain organic pigments can be for-
mulated with more durable inerts, the one most useful being
barium sulfate, known as blanc fixe. Alumina, a light form
of aluminum hydrate, is also used to make lakes or for trans-
parency. To make Alizarine Crimson, for instance, the madder
color is laked with a major proportion of alumina. Most
organic colors used, however, are concentrated toners ....
Pastes of the inerts ground in linseed oil were applied at
.003 inch thickness to aluminum foil, allowed to dry, then
exposed for four summer months to the sun in the exposure
racks used for the pigment exposure tests. Use of the alu-
minum [foil] was a test of the stresses set up in the film on
accelerated aging and of the flexibility and adhesion. The
blanc fixe formed an exceedingly tough, durable and flexible
film. It required the least oil and retained color well. The
whiting (chalk, calcium carbonate) films behaved quite curi-
ously. The wrinkling stress was so strong that it carried the
foil with it into deep wrinkles. The asbestine [magnesium
silicate] is usable where it is deliberately desired to create a
rough textured paint in not too light a color.
Barium sulfate has all the desired properties to make a
durable oil paint havi ng "body" in combination with organic
pigments and to reduce the oil proportion considerably ....
If it is further desired to increase the drying time and hard-
ness of a paint film and at the same time impart opacity, zinc
oxide and titanium dioxide in limited amount can be incorpo-
rated. By this means the Permanent Green Light pigmented
with organic toners could be formulated with but 29% oil
and have a paint that is more brilliant and tinctorially much
stronger than its inorganic counterpart. The change in color
of the organic green on drying was one-fourth that of the
. .
morgamc green.
224 Chapter 6
It was shown ... that in preparing the panels for test in
the oil vehicle most inorganic pigments could not be diluted
nearly as far as organic pigments and have a usable tint
strength. On a weight basis organic reds can be eight times
as strong as cadmiums. That means for a paint of useful
strength for artists' use, the organic color pigment content
can not only be greatly reduced with an inert, but [this] is
frequently necessary so that in mixing the color strength will
not be overpowering.
As a result, organic pigments that have a reasonable
cost can replace similar hues of inorganic pigments at a lower
price per tube. Even many of the moderately expensive and a
few of the very expensive organic pigments can be made into
artists' colors within the normal price range.
As a brief introduction, Levison's description of the use of
extenders leaves little to be added, except to point out their use in sub-
strates of paintings. For example, the grounds of commercially prepared
canvases are often treated with formulations containing inerts such as
chalk, gypsum, and barytes.
When using the Kubeika-Munk equation to calculate col-
orant formulations with pigment concentrates that might contain exten-
ders, the pigment calibration data are treated quantitatively as if the
extender pigment were a part of the vehicle system, since the refractive
indices are similar. If this assumption seems not to be completely success-
ful, a correction factor for the prepared sample in question relative to the
same pigment, albeit previously calibrated in a different vehicle system,
may be determined. (The use of a correction factor, in effect, allows one
to convert the master set of K's and S's for the colorants, determined in
a reference vehicle system, to K's and S's that will yield accurate color
matches and formulations in a particular, different vehicle system. )
When an extender pigment has been used to prepare the sub-
strate material, its presence can be ignored if hiding is complete. If the
hiding of the paint film is incomplete, so that the substrate is visible or
partially visible, the hyperbolic form of the Kubeika- Munk equation
must be used (see the sections on glazes and opacity in chap. 4 ).
In Levison's discussion quoted above, he mentions the addi-
tion of zinc oxide and titanium dioxide to an organic green pigment
(probably phthalocyanine green). Neither of these whites is an extender;
they are white pigments with significant scattering characteristics. In
essence, such an addition changes the low-scattering phthalocyanine
green toner, a t ype II pigment, into a type III pigment (see definitions
in chap. 3 ), which both absorbs and scatters significantly. A modified
organic green of this type would no longer be appropriate for use as a
transparent glaze.
Chapter 7
Reflectance Curves of Some Frequently
Encountered Chromatic Pigments
This chapter is included to comment on the salient characteristics of
reflectance curves for additional pigments; these curves may prove useful
for pigment identification. To a limited extent, an attempt has been made
to include references to a number of traditional pigments that may have
been used in older objects, as well as to a few modern pigments that may
be encountered. Unfortunately, this exposition can be only a brief intro-
duction, particularly to the many modern synthetic organic pigments as
well as to the many modern synthetic inorganic pigments, which are
often tailored to have specific reflection characteristics.
Presented are spectrophotometric reflectance curves of a
greater variety of pigments than are described in appendix A. The princi-
pal curves, all based on mixtures with rutile Ti02, are reproduced from
the article by Johnston (1967b). They illustrate a collection of recorded
and published pigments that, regrettably, encompasses only a fraction of
the vast number of pigment reflectance curves that may be encountered.
But until a more comprehensive collection of spectral curves is cata-
logued-and we hope that such a project can be carried out for future
reference-these curves offer a brief survey of pigments that may be
encountered, particularly from the years after World War II. The discus-
sion also emphasizes and enlarges upon the previously described tech-
niques of spectral curve analysis as an aid in pigment identification.
Primary Colors
The primary colors are described in the following order: blue, red, and
yellow, considering wavelengths in the order of long to short.
Blues
Figure 7.1 illustrates a series of blue pigments. The familiar ultramarine
blue (see Plesters 1993) and phthalocyanine blue are included as curves A
and B. Curve C shows a typical iron blue (Prussian blue); this was widely
used for more than two centuries but is in more limited use today. The
more or less continuous drop in reflectance in the long-wavelength red
region is characteristic of this pigment.
Curves D-1 and D-2 are both cobalt pigments. D-1, labeled
"red shade," is cobalt alumina oxide, which is the pigment generally
226
Figure 7.1
Spectral reflectance curves of paints con-
taining various blue pigments in mixture
with rutile Ti0
2
(after Johnston 1967b;
Johnston and Feller 1963).
Chapter 7
90
B
80
70
Q)
60
"' :::J

2
Q)
u
c
.lS
50
u
Q)
q::

.
+"' .
c .
Q)

40 Q)
CL
30
20
10
400 450 500
y 0 -----R----
A. ultramarine blue
B. phthalocyanine blue
C. iron blue
D-1 . cobalt blue-red shade
D-2. cobalt blue-green shade
'
'
E. lndanthrone Blue
F. indigo
'
'
'
D-1
I
', .... r ....
...._ E --
, ___ _...
550
Wavelength (nm)
600
.
.. . . . . . , .....
.. / .. )
650 700
referred to in artists' paints as cobalt blue. Curve D-2, labeled "green
shade," is a cobalt-tin oxide generally referred to in artists' paints as
cerulean blue. As can be seen, the curves of both pigments are character-
istic. Both exhibit a sharp rise in reflectance in the red region, and both
have absorption bands in the orange and red regions. The green shade
(cerulean) has bands at 645 and 595 nm; the red shade has its major
absorption at 590 nm, with another at about 620 nm. Both have slight
bands in the middle of the spectrum, cerulean in the yellow-green region
at about 555 nm, and the red shade in the green region at about 540 nm.
These cobalt blues represent one class of pigments that have a charac-
teristic curve shape in the region of maximum reflectance, providing
another identifying clue. However, if they are mixed with a pigment that
absorbs in the blue and violet short-wavelength region, such as a yellow
or orange, this characteristic shape may be obscured.
Reflectance Curves of Some Frequently Encountered Chromatic Pigments 227
Curve E, which is Indanthrone Blue (see also fig. 4.30), has
only one absorption band in the orange region and exhibits gently rising
reflectance in the red. Note that this red tail is not as great as that for
ultramarine blue, as shown in figure 4.30, and is much less steep than the
red tails of the cobalt blues. There is nothing readily characteristic about
the curve for this blue, except in terms of pigments it cannot represent.
Curve F shows a typical reflectance curve for indigo. Its
flatness relative to most blues describes its low chroma. It has an absorp-
tion maximum in the red region at about 660 nm, rising then in a little
tail in the longer-wavelength region. Because it is a natural pigment,
slight variations may be encountered, but it is generally recognizable as
indigo. As this pigment fades under light exposure, it tends to become
greener in hue. See, for example, the curve for indigo (called ai in
Japanese) found in Japanese prints (Feller, Curran, and Bailie 1984).
In the same article, the reflectance curve for aigami (dayflower blue) is
shown to have its own highly characteristic series of absorption bands,
unlike any other blue to our knowledge.
A modern blue not illustrated here, but one that could readily
be encountered in modern art, is manganese blue. Staniforth (1985) illus-
trates a reflectance curve of this blue. In older objects, smalt, a pigment
made by grinding of blue glass, may be encountered. The color is due to
the presence of cobalt, and smalt could therefore be confused with one of
the many fired cobalt pigments available today. Staniforth illustrates a
curve for smalt, as do Miihlethaler and Thissen ( 199 3). Staniforth also
shows a reflectance curve for azurite, which is a natural blue pigment
(basic copper carbonate) used for centuries. It exhibits rather flat reflec-
tance in the 610-700 nm region. A curve for azurite is also given by
Gettens and FitzHugh (1993a). The blue pigments listed in this para-
graph, but not illustrated by the reflectance curves of figure 7.1, contain
metals detectable by X-ray fluorescence. This information can be helpful
for identification.
Reds
No other group of pigments can serve as well as the reds to emphasize
the need for an organized categorization of absorption bands, particu-
larly for the many organic reds developed in the twentieth century or
used in pre-twentieth-century classical periods. In figure 7.2, the reds
exhibiting a single absorption band include the inorganic reds (the iron
oxides-curves 3 and 14; the cadmiums-curves 1 and 2) and the organic
reds (see also fig. 4.5, classical alizarin lake). Of the prevalent modern
reds exhibiting a single absorption band are some toluidines (curve 9),
Li tho! Rubine (curve 13 ), and BON red (curve 12)-all inexpensive,
high-chroma reds of great tinting strength. (A few of the other modern
organic reds, in certain media, also exhibit a single absorption band.)
Many of the organic reds exhibit two or more characteristic
absorption bands, as illustrated in figure 7.2. The number of pigments
illustrated is very small compared with the number that has been or is still
available. Of the traditional organic reds, one used by artists is carmine
lake, which is discussed by Schweppe and Roosen-Runge (1986) in Artists'
228
Figure 7.2
Spectral reflectance curves of paints con-
taining various red pigments in mixture
with rutile Ti0
2
(after Johnston 1967b).
Chapter 7
-v B
100
1. cadmium red light
2 cadmium red medium
90
3. Indian red oxide
4. Quinacridone Red
5. Pyrazolone Red
80
6. Perylene (R-6300)
7. Pyranthrone
8. Perylene (R-6500)
9. Toluidine Red
70
10. Permanent Red 28
11. Thioindigo Red

12. BON red
:::i
60
13. Lithol Rubine

14 synthetic red oxide

0)
u
c
cu
.....
u
50
0)
;;:::

.....
c
0)

40 0)
CL
30
20
10
400 450 500
G
y
550
Wavelength (nm)
0 R
/
600 650
Pigments. The absorption curve for carmine is illustrated by curve 4 in
figure 4.19. Note that it exhibits two to three absorption bands.
700
Of the inorganic reds, the one class of pigments that exhibits
a characteristic fingerprint in many situations is that of iron oxide, two
forms of which are shown in figure 7.2. Curve 3, Indian red oxide, is a
bluish version (natural); curve 14 is a synthetic red oxide, a more-orange
pigment that is also illustrated in figure A.12. The curving, rising, long-
wavelength tail in both curves is characteristic of iron oxides. We see it
again, though it is less pronounced, in the yellow oxide-yellow ocher
family, as well as in iron oxides of near-neutral hue, where it is not quite
so prominent but still apparent. See also the curve for burnt umber (see
fig. A.6). Throughout history, the iron oxides have been staples used in
coloring objects of many types. If they are not mixed with a green or
blue, which absorb in the long-wavelength region, their curved red
reflectance tail is an important aspect to note.
Reflectance Curves of Some Frequently Encountered Chromatic Pigments 229
The other classical red pigment used by artists since ancient
times is vermilion, illustrated in figure 4.6 (see also Gettens, Feller, and
Chase 1993).
A more modern red inorganic pigment extensively used by
artists since the early twentieth century is cadmium red, which has
largel y replaced vermilion (Gettens and Stout 1942). The only character-
istic in the absorption region is its flatness. The wavelength at which the
reflectance begins its steep rise in the orange and red regions describes
the hue. Cadmium red light (curve 1) and cadmium red medium (curve 2)
are illustrated in figure 7.2. Note that curve 1 rises at a shorter wave-
length than does curve 2. In Artists' Pigments, Fiedler and Bayard (1986)
provide a series of curves illustrating the masstone reflectance curves for
the family of cadmium pigments, from the yellows to the maroons.
Many of the modern organic red pigments have distinctive
multiple absorption bands, as illustrated in figure 7.2: the perylenes
(curves 6 and 8), Pyranthrone (curve 7), and the thioindigos, one of
which is illustrated by curve 11. Appendix D presents a series of red pig-
ment reflectance curves organized in a proposed orderly manner.
Yellows
Because yellows absorb in the short-wavelength violet region, there is
not much opportunity to fingerprint their absorption characteristics
(they reflect all other wavelengths) unless measurements are made in the
ultraviolet spectral region, where they may also absorb selectively. One
distinctive organic yellow illustrated in figure 7.3 is Flavanthrone Yellow
(curve 5), with its double absorption band; this modern pigment is a
very stable transparent yellow. Another recently introduced family of
yellow colors of good lightfastness (not shown here) are the azome-
thines, which are metallic complexes that also have very characteristic
absorption bands in the violet region. Although identified by their
Colour Index name as Pigment Yellow 129, they are definitely greenish.
An isoindolinone cobalt complex has three characteristic absorption
bands in the violet region (no Colour Index name), but it is definitely
more yellow in hue than the azomethines. Made by Ciba Specialty
Chemicals (formerly Ciba-Geigy), these two chemical types are all sold
under the trade name Irgazin.
A number of inorganic complexes in the yellow family are
also offered by Ciba Specialty Chemicals; they were made by the former
Drakenfeld Company, which is now a part of Ciba. Many do not have
characteristic absorption in the violet region, but they do exhibit absorp-
tions in the red and near-infrared regions. These pigments are extremely
stable to light and heat (they are fired materials) and may be found in
ceramic objects as well as in organic media. They are relatively low in
tinting strength, so they are selected primarily on the basis of their excel-
lent durability.
Some traditional yellows are cobalt yellow (Aureolin), Indian
yellow, lead-tin yellow, and lead antimonate. The only pigment of these
four that has striking characteristics in its reflectance curve in the visible
region is Indian yellow, which fluoresces strongly. It is described at some
230
Figure 7.3
Spectral reflectance curves of paints con-
taining various yellow pigments in mixture
with rutile Ti0
2
(after Johnston 1967b).
Chapter 7
90
80
70
~
60
:J
:t::
~
ClJ
u
c
.:::l
50
u
ClJ
/11"
""'
.
~
.
.
...,
.
c .
ClJ
.
u
2 .
.
cu 40 .
CL
30
20
10
400 450 500
y 0 ----- R----
T ~
5
-- 11
...................................... 4
/./
9
chrome yellow medium
primrose chrome yellow
chrome yellow light
4. synthetic yellow oxide
5. Flavanthrone Yellow
6. Hansa Yellow G
7. Benzidine Yellow
8. Anthrapyrimidine Yellow
9. Gold Paste
10. Toluidine Yellow
11. cadmium yellow light
550
Wavelength (nm)
600 650 700
length by Baer and coworkers (1986) and discussed earlier in this mono-
graph in the section on fluorescence. The others exhibit no characteristic
absorption in the violet region. Reflectance curves for cobalt yellow are
illustrated by Cornman (1986) in volume 1 of Artists' Pigments and for
lead antimonate by Wainwright, Taylor, and Harley (1986); reflection
curves for lead-tin yellow are illustrated by Kiihn (1 993a) in volume 2.
Another traditional yellow colorant used by artists is
gamboge, which is highly fugitive. It also has no striking absorption
characteristics.
Synthetic yellow iron oxide (curve 4 in fig. 7.3) exhibits the
upward curve in the long-wavelength (red) region. The curve shape is
more subtle in this region than that of the red oxides, but it is not flat
like those of the high-chroma yellows. A further characteristic of yellow
oxide is the slight dip, surprisingly persistent, at about 470 nm.
Reflectance Curves of Some Frequently Encountered Chromatic Pigments 231
One of the major clues to identifying yellows, oranges, and
reds is the wavelength at which the reflectance begins to rise sharply. It
describes the hue: the longer the wavelength, the more orange the hue;
the shorter the wavelength, the greener the hue. Compare the curves for
the three chrome pigments in figure 7.3: Chrome yellow medium (curve
1) is the most orange; chrome yellow light (curve 3) is a true yellow, nei-
ther orangish nor greenish; and primrose chrome yellow (curve 2) is per-
ceived as definitely greenish.
Secondary Colors
The secondary colors are those that may be made by the mixture of two
primary colorants: green by mixing blue and yellow, purple (sometimes
called violet) by mixing blue and red, and orange by mixing red and yel-
low. These colors may also be achieved with single compounds, not mix-
tures. The secondary colors illustrated in figure 7.4 (the greens), figure
7.5 (the purples or violets), and figure 7.6 (the oranges) are all single
compounds, with the exception of chrome green, which is a mixture of
iron blue and chrome yellow. It should be readily apparent why the pri-
mary colors should be analyzed first to alert the researcher to the possi-
bility that the secondary colors may be mixtures of the primaries rather
than single, distinct pigments. The reader is referred to the mixture
curves in appendix A: mixed greens in figure A.24, mixed violets
(purples) in figure A.27, and mixed oranges in figure A.30.
Greens
The greens are illustrated in figure 7.4. Chrome green (curve 1) has the dis-
tinct absorption in the red region that is characteristic of iron blue (curve
C in fig. 7.1 ); it is a mixture of a chrome yellow and an iron (Prussian)
blue. However, since it is traditionally sold as a separate pigment and is
given a separate Colour Index name, PG 15, it has been included here in
the green category. It may be made by admixture of dry pigments or by
coprecipitation. For our purposes, it is immaterial that it be identified as a
distinct pigment, since the properties and curves are those of the mixture.
Chrome green should not be confused with the other
chromium-containing greens: hydrated chrome oxide (curve 7), which is
the artists' Viridian or Vert Emeraude Transparent, and chrome oxide
green (curve 2), which is a fired, pure chromium oxide, low in chroma
but extremely durable, and known since the early nineteenth century
(sometimes referred to by artists as Vert Emeraude Dull). Mixed with a
neutral white or gray, each green has its distinctive absorption charac-
teristics: the double absorption bands at 465 nm and 600 nm of the pure
chrome oxide are unmistakable. The major absorption of the hydrated
chrome oxide at about 625 nm could be confused in shape with the
curves of blues, but note that the absorption maximum is at a longer
wavelength where greens absorb. One unique characteristic of the
hydrated chrome oxide, if it is present without admixture with a yellow,
is the little absorption dip in the high-reflection region, at about 435 nm.
232
Figure 7.4
Spectral reflectance curves of paints con-
taining various green pigments in mixture
with rutile Ti0
2
(after Johnston 1967b).
Chapter 7
-V -*4-- B - -->+4--- G y 0 -----R----
100
90
80
70
60
50
40
30
1. chrome green
2. chrome oxide green
3. tungstate green toner
4. Pigment Green B
5. Green Gold (Nickel Azo Yel low)
6. phthalo green
7. hydrated chrome oxide
/"'
./ \
/ .
. \
/ - ---,-- /
7 v - -s- /
/ //.. ___ ....
/"- : I \ . . \
I
. I I ': .. \ . )
. I . /
. 1' I '\. / .
/ . . , ..
2 /: I ? . ...._ _ _;:/
/ . I 2
j.' i .
i'
.
1:
,.
..
I
I
.
.
20 I
l
10
400 450 500 550
Wavelength (nm)
600 650 700
Of the other greens illustrated in figure 7.4, phthalocyanine
green (curve 6) is by far the most important green pigment used since
about 1950. It is without doubt the most prevalent modern green pig-
ment. Its distinctive tai l in the red region is characteristic.
Curve 4 in fi gure 7.4 is the reflectance curve for Pigment
Green B (Pigment Green 8 ). Since the early twentieth century, the artists'
color known as Hooker's Green may have been made with this pigment.
According to Levison (1976), Hooker's Green was originally made wi th
iron blue and gamboge. One can assume that since the gamboge faded so
rapidly relati ve to iron blue, trees in some old artworks have become
blue. The only characteristic of the Pigment Green B reflectance curve is
the very flat absorption in the violet region, 415-440 nm. In the red
region, no maximum absorption is exhibited. The dropping off of
reflectance is reminiscent of the curve of an iron blue, and, indeed, this
pigment is an iron chelate. The chroma is low, with a skewed (nonsym-
Reflectance Curves of Some Frequently Encountered Chromatic Pigments 233
metrical) reflectance curve. It is still used in watercolors and in polymer-
emulsion artists' colors. Unless used singly, it may be difficult to identify
this pigment positively from reflectance alone.
Not included in figure 7.4 are several classical green pigments
that are found in older works. One of these is Green Earth, whose
absorption curve is illustrated in figure 4.9. In Artists' Pigments, volume
1, Carol Grissom (1986) discusses its spectral reflectance at some length.
The characteristic to be emphasized here is the continuous drop in
reflectance in the red region; this could be confused with iron blue's char-
acteristic drop in reflectance in the same region.
Other classical greens are verdigris, which is generally dibasic
acetate of copper, and copper resinate, based on verdigris dissolved in a
resinous solution. Malachite, which is the natural mineral of basic
copper carbonate, and green verditer may also be encountered. Kiihn
(1993b) has written about and shown reflectance curves for verdigris
and copper resinate. Gettens and FitzHugh (1993b) illustrate reflectance
curves for malachite and green verditer. Curves of some mi xtures of mod-
ern blues and yellows are illustrated in figure A.24.
Purples
The purples are sometimes referred to as violets. In Munsell notation,
they are designated by the letter P; in the Colour Index, by the letter V.
We have chosen to call them purples (P) to emphasize that a similar hue
may be obtained by mixing blue and red.
The reader is again referred to the purples made by the mix-
ing of blue and red illustrated in figure A.27.
The reflectance curves of the purples (violets) illustrated in
figure 7.5 are highly characteristic; they are all organic in nature, recent
twentieth-century developments. Each has a characteristic number and
location of its absorption bands, except alizarin maroon. Not included in
the figure, but commonly used by artists in the twentieth century, is man-
ganese violet. Introduced at the turn of the century, it also has a unique
absorption curve (Patton 1973).
A violet (purple) pigment that might be encountered in art-
works is cobalt violet, generally cobalt phosphate (or the arsenate, not
generally used because of its toxicity). Its reflectance curve would possi-
bly display the multiple absorption bands characteristic of the cobalt
blues, with the absorption peaks shifted to longer wavelengths. No
reflectance curve for this pigment has been found.
Oranges
A series of oranges that are not mixtures of reds and yellows is
illustrated in figure 7.6. Of those illustrated, only the Halogenated
Anthanthrone (curve 1) has a striking fingerprint . The others are more or
less flat through the blue region and into the green. The wavelength at
which the rise in reflectance occurs determines the hue-the longer the
wavelength of the rise, the redder the hue. Many of these oranges can be
closely matched by mixtures. For example, cadmium orange (curve 7)
could probably be made by admixture of the right pair of cadmium
234
Figure 7.5
Spectral reflectance curves of paints con-
taining various purple pigments in mixture
with rutile Ti0
2
(after Johnston 1967b).
Chapter 7
y 0 -----R----
1. Quinacridone Violet
2. Thi oindigo Red
90 3. Alizarin Maroon
80
70
60
4. carbazole dioxazine
5. lsoviolanthrone Violet
.. -..
; '
I '
I '
\
\
2 - -
.
I
.
I
.
I
50
,
,
,
,
4
,'I /
40
30
20
10
400 450
.
;
/I l
, /
\\ _,.-- I /
3 /9 .
~ \ 1 1-.......1
500
: : : : : : : ~
550
Wavelength (nm)
600 650 700
yellow and cadmium red pigments. Figure A.30 shows the reflectance
curves of three oranges made by mixture of a red and a yellow pigment.
Pigment Interactions
There are few interactions among modern pigments. (There are restric-
tions due to the nature of the medium or vehicle used, and such informa-
tion is generally supplied by the pigment manufacturer or supplier.)
There is, however, one striking exception: phthalocyanine blue appears to
"dissolve" in at least one pigment-notably with Benzidine Yellow, one
of the diarylide yellows, in which phthalocyanine blue appears to form a
solid solution. DiBernardo and Resnick (1959) first pointed out this phe-
nomenon. Figure 7.7 is an illustration of spectral curves of a phthalo
blue and a Benzidine Yellow in a textile printing vehicle. The illustration
Figure 7.6
Spectral reflectance curves of paints con-
taining various orange pigments in mixture
with rutile Ti0
2
(after Johnston 1967).
Reflectance Curves of Some Frequently Encountered Chromatic Pigments

::>


Q)
u
c:
m
t
Q)
"" [!:'

c:
Q)
u
Q;
0...
1. Halogenated Anthanthrone Orange
2. Dianisidine Orange
90 3. Pyrazolone Orange
4. molybdate orange
5. chrome orange
6. Benzidine Orange
80 7. cadmium orange
8. Orthonitraniline Orange
9. Dinitraniline Orange
70
60
50
40
30
20
10
400 450 500 550
Wavelength (nm)
600
235
650 700
for each pigment shows the normal reflectance curve. Also illustrated is
the curve of the combination of the two pigments in the same vehicle. It
can be seen that the phthalo blue absorption in the long-wavelength
region is very different from the normal curve of phthalo blue illustrated
in figures A.10 and 7.1. This "solution" curve occurs as well in the pres-
ence of aromaticity in vehicles. For example, the absorption shape begin-
ning at about 540 nm is the only characteristic absorption observed in
polystyrene plastics. Many paint vehicles, particularly those containing
aromatic groups or involving aromatic solvents, may exhibit curves that
dip at about 680 nm, indicating a small amount of phthalocyanine blue
in this "soluble" form. (Phthalo blue can often be observed as a contami-
nant in colors with high reflectance in the long-wavelength region. See
the discussion under glazes in chap. 4. )
As mentioned, such interaction between modern pigments is
unusual. Most pigments today are very stable and unreactive. However,
236
Figure 7.7
Spectral reflectance curve of a phthalo
blue and Benzidine Yellow in a textile
printing vehicle. The reflectances of the
individual pigments in the same vehicle are
also shown. The change in reflectance in
the long-wavelength region from the typi-
cal phthalo blue absorption is characteristic
of the absorption of phthalo blue in aro-
matic paint systems.
Chapter 7
100
90
80
70
60
<!)
u
c
_:g
u
<!)
50
;:;::
[':'
!il
0
I-
40
30
20
10
400 450 500
G y 0
-----R----
(2) Benzidine Yellow
(3) mixture of 1 and 2
(1) phthalocyanine blue
550
Wavelength (nm)
600 650 700
they may fade, and they may be unsuitable for certain media. For
example, carbazole dioxazine violet is extremely sensitive to peroxides.
However, in general, each pigment's unique reflectance curve is not
altered by the presence of another, except as would be predicted by
overlapping absorptions and reflectances according to the Kubelka-
Munk predictions.
This brief summary of reflectance curve characteristics of vari-
ous pigments can only suggest the information that can be learned from
experience and from an extensive reference library of pigment reflectance
curves. There is no substitute for learning from the study and measure-
ment of many known pigments. There is no substitute for concentrated
effort. There is no substitute for a knowledge of the available pigments
and their properties. Of course, the same criteria apply to almost all non-
routine analytical applications of any nondestructive reflection technique,
whether it be wavelength dispersive spectrometry in the X-ray region (for
Reflectance Curves of Some Frequently Encountered Chromatic Pigments 237
metals), in the visible region as described here, in the near-infrared region,
in the mid-infrared region, or in the ultraviolet region. The same stan-
dards of analytical spectroscopy in the classical sense apply: careful
instrument calibration, reliable reference standards, a reference library of
known materials, recognition of variant results, and standard analytical
knowledge of sensitivity and precision. Most of such expertise is achieved
from extensive experience and dedicated continuing study.
Special Technique for Dark or High-Chroma Colors
When colors are very dark, such as a navy blue, or are very high in
chroma, such as a bright orange, the characteristic absorption bands are
not revealed because the reflectance is very low in the absorption region.
Thus, the measurement of the reflectance does not yield the information
needed for identification.
If a very small sample can be taken, a dilution in white paint
can be made to bring out the characteristic absorption bands. The tech-
nique consists of dispersing the very small sample either into white pig-
ment from which paint can be made, or directly into wet white paint.
This dispersion can be carried out with a micro ball mill such as is used
for grinding micro samples for other analytical methods. McCrone's
Micronizing Mill is an example of such a device. A drawdown or paint-
out of the white dilution paint is made and dried, and then its reflectance
is measured for comparison with the library of curves in the normal way.
Although this method requires that a small sample be taken, the sample
is not lost-it still exists to be used in solution spectrophotometry or
another appropriate analytical procedure. The technique of diluting a
dark sample by dispersing it in mixture with a white pigment has proved
to be a worthwhile supplement to the arsenal of analytical methods use-
ful for identifying colorants.
Chapter 8
Measured-Data Analysis and Special Measurement Problems
In this chapter a few topics relevant to color measurement and colorime-
try, not previously emphasized, are discussed briefly. In general the topics
involve the interpretation of multiple measurements, either those made
over a short period of time or those made over periods of many months
or years. Of particular emphasis is the importance of making multiple
measurements on nonhomogeneous or mottled areas. The major points
to be made are as follows: (1) Know the reliability of the measurements
made on your instrument by following a procedure of calibration and
measurement mentioned in chapter 1 (see Hemmendinger 1983 ). (2)
Look at the spectral curve shapes of the measured samples and analyze
the nature of any differences. (3) Look at the samples. (4) Plot the mea-
sured chromaticity coordinates on a chromaticity diagram. Think about
what each of these observations is telling you. Analyze- so the total pic-
ture can be constructed mentall y.
Note that there is no mention above of averaging tristimulus
values or chromaticity coordinates. It must be remembered that tristimu-
lus values are dependent variables and so should not be averaged. In view
of this, how does one select the most representative value from a series
of repetitive measurements, sampling, sample preparations, and material
preparations? The answer is to plot the data on a graph and look at the
distribution. For example, plot the chromaticity coordinates on a chro-
maticity diagram, noting the Y value alongside the points (or plot x
against Y), in addition to x against y. Alternativel y, the derived uniform
chromaticity coordinates, such as a", b''., and L ", may be employed. On
the basis of the plotted data, select the center of the distribution as the
best average value. The important point to be emphasized is that one
must not simply average thoughtlessly. By doing so, one could get a very
false picture or a set of figures that might never even be possible with the
particular set of materials being used.
The following illustration, utilizing the concepts outlined in
the first paragraph, should help to make the point more lucidly than
words alone could do. A new color was being introduced in a polyester
gel coat of moderate gloss. It was a turquoise made with 4.9% phthalo
green, 0.01 % black, 1 % yellow iron oxide, and 94.09% Ti02 The prob-
lem was to select the best CIE coordinates to use as the standard in
preparing subsequent batches. Table 8.1 shows the data measured on six
panel preparations from the same batch of material.
Table 8.1
Panel preparations of turquoise gel coat
(illuminant C, 1931 standard observer).
Figure 8.1
Greatly enlarged portion of the Cl E 1931
chromaticity diagram, based on data from
table 8.1, showing the locations of points,
with corresponding Y values, for the six
turquoise gel coats, all made from the
same batch of original material. The
496 nm dominant wavelength line is
shown. Sample 6 was selected as the best
standard to use for future preparations.
The formulation contained a little yellow
oxide, which could be withheld in subse-
quent manufacture to allow for a vehicle
that might be more yellow. Other compo-
nents were phthalocyanine green, a little
black, and rutile Ti0
2
.
Measured-Data Analysis and Special Measurement Problems 239
CIE Sample number in order of preparation
coordinates 2 3 4 5 6
x 0.2263 0 .2276 0.2285 0.2272 0.2247 0 .2259
y 0.3483 0.3517 0.3529 0.3484 0.3494 0 .3489
y
31.47 30.94 30.90 31.65 30.79 31.21
All of these measurements were made with the specular
reflectance excluded. Three areas of each panel were measured; the pan-
els were very uniform, with negligible deviations over their areas.
Figure 8.1 shows the chromaticity val ues plotted on a portion
of the chromaticity diagram. This turquoise is in the green-blue region,
where yellowing is very objectionable, with a dominant wavelength
between 495 nm and 496 nm.
After the first three panels were made, a red flag was raised
because visual inspection of the panels and their measurements indicated
that samples 2 and 3 were increasingly yellow, possibly because of resin
yellowing caused by excessive heat. The technician preparing the panels
corrected the problem and prepared three more panels, samples 4, 5, and
6. Visual examination of all of the panels suggested that there were slight
differences in gloss. Examination of the spectral curves indicated that to
be the case. Sample 6 was selected as the sample that was closest to the
"normal color," and it was chosen as the best physical standard to use for
future preparations. This was in keeping with a policy of selecting mea-
surements made on a real sample close to the center of the distribution.
0.354 ....---------------------------.
0.352
0.350
0.348
0.346
0.222
y = 30.79

5
Y=31.21

1
Y=31.47
e measured data
overall average
y = 30.90

y = 30.94

2
3
e Y=31.65
4
Q selected best" average" sample
0.224 0.226 0.228
x
0.230
240 Chapter 8
Samples 2 and 3 were rejected as clearly nonrepresentative. Gloss differ-
ences in medium-gloss samples, which have somewhat diffuse surfaces, are
difficult to control. In general, however, variations of this type are less
objectionable than hue differences. In particular, hue differences in the
yellow direction in a turquoise are seriously objectionable. In addition,
selecting panel 6 provided a real sample to keep as a future reference
standard. After subsequent preparations of different batches, the standard
could be changed slightly, if that appeared warranted.
Although this is an example from an application in industry,
it serves to illustrate the points made in the initial paragraph: look at the
spectral curves; look at the samples; plot the data; and think about what
all of these observations are telling you. Look and think-the admonition
by Billmeyer and Saltzman (1981) presented here in reverse order-sums
up the process well. One cannot overemphasize the value of plotting the
measured color coordinates and looking at the samples and data, and
then thinking about the implications. This is especially true when
changes in color are analyzed following exposure (Johnston-Feller and
Osmer 1977).
The differences measured over long periods of time must be
analyzed in terms of the instrument's repeatability over the same length
of time. When changes are small and of the order of magnitude of the
instrumental uncertainties, plotting the data can be helpful. One looks
for trends: If, in light of the measurement uncertainty, a difference occurs
that is within the uncertainty but that in later measurements continues to
occur in the same direction, then credence can begin to be placed in the
validity of the change.
Does this mean that averaging should not be done? Generally,
averaging is improper, as has been pointed out, but it can be used in
certain cases. An average is as representative as any other single value
( 1) when the number of measurements is large and the differences are
small-for example, in the evaluation of the repeatability of measurements;
(2) when the chroma, or saturation, of the colors is very low; and (3) when
the variation in the measurements is large because of the nature of the
samples, and the number of measurements is also large. Here averaging
may be necessary to take into account the large variations in the samples.
An example of this situation is that of measuring the color of
coffee beans. A large bag of beans sent to the laboratory was first sorted
visually into four groups based on their lightness. Measurements were
then made by placing the beans in the lid of a petri dish, about 10 cm
(4 in.) in diameter, with the shiny side of each bean against the lid. When
the lid was covered, the dish was filled tightly with additional beans and
taped shut. The lid side was placed at the port of the sphere on the Trilac
spectrophotometer, and measurements were made with the specular
reflectance excluded. The dish was rotated 90 and another measurement
made; this was repeated until four measurements were taken. The
process was repeated with another selection of beans from the pile.
Finally, the petri dish was filled with no orientation of the beans, and
five more measurements were made as the petri dish was rotated. A total
of thirteen measurements was thus made on the darkest beans.
Figure 8.2
Typical spectrophotometric curves of the
four groups of visually sorted coffee beans,
which were packed into petri dishes and
measured with their dark sides facing the
Trilac spectrophotometer.
Measured-Data Analysis and Special Measurement Problems 241
The process was repeated on each pile of beans, although the
number of fillings was reduced to one with the beans oriented and one
with them randomly oriented, each measured four times and averaged.
Figure 8.2 shows typical reflectance curves measured for each of the four
categories. Note that the reflectances have been multiplied by 5. There is
nothing particularly distinctive about the curves.
In figure 8.3 the chromaticity coordinates for all of the
measurements are plotted. The spread in the data is so large, and the
reflectance curves are so similar in character, that averages were calculated
for each group. Selecting a specific measurement in the center of the dis-
tribution would have served just as well, as in the case of the turquoise
gel coat. The important point is that one knows much more about the
data by plotting and looking at the results. In this situation, every mea-
surement was in one sense an average made by the spectrophotometer.

::::!
:;:
;s
QJ
u
c
.L'l
u
QJ



c
QJ
u
Q::;
Cl..
70
60
50
40
x5
darkest
30
20 coffee beans (in petri dishes)
typical curves of each category
10
400 450 500 550
Wavelength (nm)
600 650 700
242
Figure 8.3
Portion of the CIE 1931 chromaticity dia-
gram showing the locations of the mea-
surements made on the four categories
of coffee beans. The percentage figures
on the diagram are the relative weight
amounts of the sorted beans in each cate-
gory. The average luminous reflectance is
indicated alongside the clusters of mea-
surements for each category.
Chapter 8
0.37
0.36
0.35
0.34
""
0.33
0.32
0.31
0.30
D
Illuminant C
0.31 0.32
25.6 %
0.33
Av. Y = 14.0
co
/;,. NjJ
A Y
--10.8 17.2 %
v.
"4. Jf/;,.
$ o<o
Av. Y = 8.8 32.8%
OJ;.
o}J;,.
24.4 %
0.34
x
coffee beans
O oriented
random
e average, each category
0.35 0.36 0.37 0.38 0.39
This example is presented for another reason. There are many
times when the color of very small objects must be measured; or the
color of unevenly colored areas, such as wood with its obvious grain,
must be studied. In either of these instances, multiple measurements are
essential, and plotting the derived data on a chromaticity diagram can
aid tremendousl y in their interpretation. Blindly averaging the results
does not allow for an intelligent interpretation of the measurements.
Once the sample with the best representative color coordi-
nates has been selected as the standard, the total color differences, t:..E,
from this chosen standard can be calculated. The total color difference is
a deviation from the adopted standard and can be treated by standard
statistical techniques. For example, in order to determine the significance
of measurements, it is useful to calculate the standard deviation, cr or s.
It must be remembered t hat deviations cannot be averaged to determine
their statistical significance. The deviations must be converted to vari -
ances, the variance being the square of the deviation. The variances can
then be averaged. The standard deviation is the square root of the aver-
age variance. When the number of measurements, n, is less than about
20, one less than this number, n - 1, is used for averaging. Stated as an
equation,
1
sJ l2
l n-1 J
(8.1)
In general, s represents the difference that will occur about 67% of the
time. Classically, the 2s limit, which represents about 95 % confidence, is
considered a safe limit for the establishment of specifications. Only 5%
of the time will results occur that are outside this limit.
Measured-Data Analysis and Special Measurement Problems 243
An example is the 13 measurements made on the dark coffee
beans. The standard deviation calculated using equation 8.1wass=1.63
ls= 3.16 One measurement exceeded the ls limit, having
a a,b of 3.6. This procedure was also the statistical method used in
adding up the individual color differences of each step in the testing of
batches of paints described by Johnston (1963).
The example of the coffee beans represents the problem of
measuring mottled or uneven areas, such as the samples of wood men-
tioned previously. Two approaches can be taken, depending on the
nature of the unevenness: (1) if there are patches that are uniform over
an area equal to the smallest area of viewing on your instrument, then
each of the individual colors in the mottling can be measured separately;
(l) if not, measurements can be made with the largest area of view
possible. In both cases, take as many measurements as seem necessary
to get a picture of the general, overall effect. If measurements are to
be made at a later time to determine if any change has occurred, an
arrangement must be made to record an accurate location of each area
measured. One such arrangement was described by Feller, Curran, and
Bailie (1984) for two different instruments used to measure the colors
in Japanese prints. Also see the very precise arrangement described by
Saunders and Cupitt (1993).
Another difficult color-measuring task is to try to characterize
objects that do not have flat, planar surfaces. Curved objects, such as
vases, are an example. In general, one should measure total reflectance
using the smallest area of view possible and make a number of measure-
ments. If the material has a high gloss and is a dielectric (nonmetal), cor-
rections can be made mathematically for the gloss, as previously described
in chapter 4; if it is a metal, the total reflectance is used (see chap. 5).
Many objects present difficult measurement problems when
they are nonuniform. Examples are the pointillist dots or the normal
streaks and scumbles of many areas in classical painting. This problem
was alluded to in the discussion of additive color mixture and subtractive
colorant mixture in chapter l. Just as the human eye integrates such
irregularities under certain viewing conditions, so does a color-measuring
instrument. The results can be difficult to interpret. Only by making
many measurements in the most uniform areas can a possible palette of
colorants used by the artist be identified.
Chapter 9
Instrumentation Overview: The Tasks Determine the
Selection
The early chapters in this monograph referred to characteristics of instru-
ments used in the study of color problems. The basic design features of
spectrophotometers were described in chapter 1, and the calculation of
colorimetric descriptions from measured spectrophotometric data was
explained in chapter 2. Tristimulus filter colorimeters designed primarily
for the measurement of color differences were also mentioned. Later
pages dealt with the applications of instrumental measurements of the
light reflected or transmitted by objects.
From these discussions, one major conclusion is obvious:
the nature of the problem to be solved must first be defined before an
instrument appropriate to the task can be selected. Is the problem
an analytical problem? Does one wish to measure spectral curves for
identification of the colorants used? Does one wish to calculate colorant
concentrations for color matching, or to anal yze the changes in spectral
curves following exposure over periods of time? These are examples of
analytical problems. Throughout, it has been emphasized that a spec-
trophotometric curve in itself is not a description of the color observed.
The science of colorimetry derives from the spectral data of the illumi-
nant and standard observers used, as well as from the measured spectral
data of the object to be described. Spectral data describing the illumi -
nant and standard observer color-matching functions are necessary to
reduce spectral data to terms that may be related to what we see as
color. It has been emphasized that both treatments of spectral data-
colorimetric and spectrophotometric-are useful and complementary.
Thus, three-dimensional colorimetric data, such as tristimulus values,
are not very useful in solving analytical problems, and spectral data
alone are not generally useful in describing color or color differences.
Perhaps it is not necessary to point out that colorimetric data may be
obtained from spectral data, but spectral data cannot be determined
from colorimetric data.
Spectrophotometry is the fundamental tool used in the study
of color. The sensitivities of human visual receptors are described in spec-
tral terms; the sensitivities of other photodetectors, such as the detector
used in an instrument, the film in a camera, or the phosphors in a televi-
sion tube, are also described in spectral terms; so is the light source. And
the appearance of the object, the primary aspect discussed in this mono-
graph, is characterized by its spectral reflectance or transmittance.
Instrumentation Overview: The Tasks Determine the Selection 245
Spectrophotometry is not, however, an exact science: spectral
data depend on the method used to measure them. Factors affecting the
measured data depend on the design of the particular instrument used in
relation to the characteristics of the material being measured, as well as
on the skill and knowledge of the operator. These factors in combination
determine the precision of the measurements. The precision of the mea-
surements-that is, their repeatability and reproducibility-must be
known. Repeatability is here defined as the closeness of the agreement
between the results measured successively on the same test specimen by
the same operator using the same instrument (ASTM Standard E 284).
Thus, repeatability depends on the particular single instrument used, on
the skill of the operator, and on the nature of the sample. For measure-
ments made over periods of time, reproducibility is of prime importance.
Reproducibility is defined as the closeness of agreement of measurements
made over periods of time, perhaps by a different operator or by use
of a different instrument. Over long periods of time, reproducibility
approaching the ability of the human observer to detect small color dif-
ferences is very difficult to achieve. Careful instrument calibration and
a systematic and continuing program of measurement involving stable
reference standards (generally ceramics) are required even to approach
such a goal. The reference standards need to include samples in various
regions of color space and with varying degrees of spectral complexity.
These admonitions apply to tristimulus colorimetric measure-
ments as well as to spectrophotometric measurements. It is absolutely
essential that these characteristics of any analytical instrument used, that
is, the repeatability and reproducibility, be known before the significance
of any measurements can be interpreted meaningfully.
In 1971 a report of the Inter-Society Color Council's
Subcommittee for Problem 24, "Color Measuring Instruments: A Guide
to Their Selection," was first published (Johnston 1971a). It consisted of
a review of instrument characteristics to be considered when an instru-
ment is selected, and it included a basic checklist for analyzing an indi-
vidual user's problems and needs. In 1983 the report was reprinted by
the American Association of Textile Chemists and Colorists as part of
their summary publication Color Technology in the Textile Industry
(Celikiz and Kuehni 1983). Many of the basic ideas presented in this
publication are equally valid today. Of particular value is the preparation
of a checklist or summary of the purposes for which an instrument is to
be used (problems to be solved), and of the various types of instruments,
along with their characteristics, that might be used. Thus, a list of avail-
able instruments and their characteristics constitutes the second major
summary that needs to be made before a logical procedure for solving
the color problems at hand can be outlined.
One of the most significant advances in color-measuring
instruments made in recent years is the advent of portable colorimeters
and spectrophotometers. A summary of the various types was presented
by Kettler (1995a) and includes spectrophotometers with 10 nm or 20 nm
band intervals, as well as colorimeters. In addition, some "abridged"
goniospectrophotometers are described that have fixed angles of
246 Chapter 9
illumination and view, generally at three, four, or five angles. Some utilize
fiber optics for light transmission.
In a subsequent publication, Kettler (1995b) presented a sum-
mary of stationary instruments on the market at the time, including spec-
trophotometers, colorimeters, and gonio instruments of the abridged
type. One goniospectrophotometer, with variable angles of illumination
and view over 5 intervals of angle, is described. Made by Carl Zeiss
Jena, this instrument measures at 3 nm intervals of wavelength over a
320-720 nm range, using varying angles of illumination and view. Not
mentioned in Kettler's review is a goniospectrophotometer manufactured
by Murakami Color Research Laboratory in Japan and marketed in the
United States by Hunter Associates Laboratory, Incorporated. Both the
incident and the viewing angles can be changed. The Trilac goniospec-
trophotometer, described in chapter 1 and used for the variable-angle
measurements reported in chapters 5 and 6, is no longer manufactured,
but it may still be in use in some laboratories.
There are several reasons for understanding instrument char-
acteristics, in addition to the need to make a decision concerning the most
suitable instrument to purchase. If one already has a color-measuring
instrument, or has access to one, there is the need to know exactly what
information one can expect to obtain with it. Also, when studying the lit-
erature and data reported by previous workers, one must be able to assess
the precise meaning of the published data in order to correctly evaluate
their significance.
Material (Sample) Characteristics
Of basic importance in defining the problems to be solved and in select-
ing the best measuring instrument for the task is the consideration of the
nature of the materials or samples to be measured. Again, it is advisable
to make a list of the types of samples one wishes to study. Throughout
this monograph, the types of instruments necessary for measuring vari-
ous kinds of materials have been described.
If transparent samples-liquids or solids-need to be mea-
sured, an instrument capable of making transmission measurements is
necessary (see chap. 4). Many of the tristimulus colorimeters using bidi-
rectional geometry have no provisions for making such measurements. In
Kettler's lists of color-measuring instruments (Kettler 199 5 a, b ), this fea-
ture is not generally included; the specifications obtained from the manu-
facturer must be consulted.
Translucent samples (l iquids or solids) present a difficult mea-
surement problem. (See the section on opacity, translucency, and hiding
power in chap. 4.) Because such samples scatter light, an instrument
equipped with an integrating sphere that allows for both transmission
and reflectance measurements, with a provision for measuring both total
and diffuse reflectance, is most desirable.
For the measurement of opaque, uniform, dielectric (non-
metallic) materials, most of the color-measuring instruments are satisfac-
Instrumentation Overview: The Tasks Determine the Selection 247
tory, and their selection depends primarily on the purpose of the mea-
surement. If one wishes to measure the color difference of materials of
identical colorant composition and surface characteristics, tristimulus
filter colorimeters are very satisfactory. There is no need for more elabo-
rate instrumentation. An example of this type of application is quality
control in industry.
However, if one wishes to calculate colorant formulations
necessary to produce a desired color or to analyze the source of color
change (in reality, a colorant composition calculation), then the use of
spectrophotometry is essential. (See the section on absorbing and scatter-
ing materials in chap. 4.) This task would seem to be a straightforward
problem, as some marketers of "push-button" spectrophotometers and
computer color-matching systems would have you believe. Although such
systems may suffice for many industrial applications, the problems asso-
ciated with this limited approach largely stem from the small number
of wavelengths measured and used for the calculations. If the same col-
orants can be used to calculate the concentrations necessary to match a
given color, such a system may suffice. However, if the same colorants
may not be involved in matching a color with a minimum degree of
metamerism, as is the case in matching paints to a textile or a ceramic,
for example, measurements and computations based on 20 nm intervals
may not provide a satisfactory solution. (See the discussion of meta-
merism in chap. 2.) A larger number of wavelengths, such as those based
on 10 nm, 5 nm, or 2 nm intervals, is desirable.
A basic problem to be considered in formulating colorant
concentrations to match a standard is the surface reflection characteris-
tics of the reference sample relative to those inherent in the material to
be used to make the match. An instrument with an integrating sphere
and one that is capable of both including and excluding the specular sur-
face reflection can be helpful. Admittedly, however, there are times when
differences in the nature of the surface reflections are a necessary part of
the problem, so that visual judgments are often the most satisfactory
basis for making a useful compromise.
Throughout these pages, the importance of the surface
reflection of dielectrics has been emphasized. Changes in the surface
reflection characteristics must be considered, particularly when materials
must be evaluated before and after exposure. It is highly desirable to
select a color-measuring instrument equipped with an integrating sphere
that will measure both the total reflectance (specular component of the
reflectance included, SCI) and the diffuse reflectance (specular component
of the reflectance excluded, SCE).
For the measurement of the reflectance of metals-that is,
nondielectric materials-where the color comes from the reflection at the
surface, an instrument capable of measuring the total reflection is essen-
tial. If one attempts to measure the color of a highly polished, flat piece
of gold, for example, using an instrument with bidirectional geometry or
one with an integrating sphere that does not have a provision for includ-
ing the specular reflection, little or no reflectance will be measured. The
proper instrument among those generally used for color measurement is
248 Chapter 9
one that measures the total reflectance (SCI). Generally, an instrument
equipped with an integrating sphere is used. When the instrument is also
capable of excluding the specular reflection-that is, of measuring the
diffuse reflectance (SCE)-the specular reflectance (S) alone is then
obtained by the difference: S = SCI - SCE. The ratio of the diffuse
reflectance to the specular reflectance, SCE/S, provides an indication of
the degree to which the polish approaches a perfect mirror surface. (See
the section on the color of metals in chap. 5.)
Fluorescent materials have come into common usage today.
There are those with obvious fluorescence, such as the fluorescent
whitening agents (FWAs ) found in most laundry detergents, as well as
the high-visibility dyestuffs used in hunters' garb or in warning signs,
but others may also be encountered. Many dyes fluoresce; some natural
pigments contain fluorescent components; and some modern materials
may be made that incorporate fluorescent materials for other reasons.
Regardless of the source of the fluorescence, an instrument used for the
measurement of such materials must illuminate the sample before sepa-
rating the reflectances (or transmittances ) into the component wave-
lengths or tristimulus responses. Such an arrangement is generally
described as having polychromatic illumination, as opposed to mono-
chromatic illumination. Such an optical arrangement is essential to mak-
ing measurements that correlate with what one observes. (See the section
on fluorescence in chap. 6. ) When illuminating materials with the total
light from the lamp, the wavelength characteristics of the lamp emission
become important in calculating color coordinates. Many of these instru-
ments use xenon flash lamps that include ultraviolet radiation, filtered to
approximate the CIE D65 illuminant closely. ASTM Standard E 124 7,
"Standard Test Method for Identifying Fluorescence in Object-Color
Specimens by Spectrophotometry," describes techniques for detecting
fluorescence that may not be evident by visual means alone.
The use of metallic flake pigments has become widespread
in recent years; the most commonly observed use of aluminum flakes in
paints is for automotive finishes. Coated vinyl upholstery fabrics contain-
ing metallic flakes have also been used. (See the section on flake pigments
in chap. 5.) Because such flakes are relatively large and of low density,
they tend to orient primarily parallel to the surface, but they do not com-
pletely do so. Thus, these materials change in color intensity depending
on the relative orientation of the flakes, and they require goniospectro-
photometers to measure the reflectance at various angles of illumination
and view. One type of such an instrument was described in chapter 1.
Metallic flakes have been used by artists for many years, primarily for
simulating silver, gold, bronze, and copper.
Materials that exhibit pearlescence or iridescence also require
the use of a goniospectrophotometer to measure their unique color effect.
In this case, the critical reflection is specular, which must be measured at
an angle equal to and opposite the incident angle. In the use of chromatic
interference materials, the reflected color is very high in chroma. Since
many such materials are essentially transparent-that is, they scarcely
absorb incident light-the wavelengths transmitted are complementary in
Instrumentation Overview: The Tasks Determine the Selection 249
hue and very much lower in chroma. (See the discussion of pearlescence
and iridescence in chap. 5.) The transmitted color can be measured at an
angle off the specular angle provided it is reflected by a substrate.
A further application where a goni ospectrophotometer is
helpful, though less obviously so, is in the characterization of matte or
nearly diffuse reflecting surfaces. This elusive aspect was discussed in the
section on surface reflection in chapter 6.
Other aspects of the appearance of objects are likewise
important, such as the homogeneity of the color in the area to be mea-
sured. Nonuniform colored materials can be difficult to measure and
describe with certainty. If there are areas where the individual colors
can be measured with an available instrument using a very small area
of illumination and view corresponding to the size of the areas, such
an instrument should be very helpful. If not, an instrument with a very
large area of view may be of most value; valuable information can be
garnered by making many measurements over the area, then examining
the spectral curves (considering both additive color mixture and subtrac-
tive colorant mixture), calculating the tristimulus values and chromatic-
ity coordinates and plotting them. An example was presented in chapter
8 describing the measurement of coffee beans. Wood grains are another
example where the general color effect of the wood can be characterized
using this multiple-measurement technique. Thus, sample size and
instrument area of view are important characteristics to be considered in
selecting and using an appropriate color-measuring instrument. One may
wish to select an instrument that provides several viewing areas, so that
the operator can select the smallest for small samples and a larger one
for samples where appropriate.
In this monograph no mention has been made of the mea-
surement of the spectral characteristics of light sources or of the intensity
of light incident on, transmitted by, or reflected by materials in a display.
These measurements require the use of a spectroradiometer, an instru-
ment for measuring spectra of colors viewed at a distance away from the
instrument. In studying spectroradiometry, a good place to start is ASTM
Standard E 1341, "Standard Practice for Obtaining Spectroradiometric
Data from Radiant Sources for Colorimetry." In this practice, other
sources of basic methods of radiometry are listed: ( 1) Radiometric and
Photometric Characteristics of Materials and Their Measurement (CIE
1977); (2) The Spectroradiometric Measurement of Light Sources (CIE
1984); (3) "Fundamental Principles of Absolute Radiometry and the
Philosophy of the NBS Program," from the National Institute of
Standards and Technology (formerly National Bureau of Standards, NBS)
(NIST 1968-1971); and (4) Guide to Spectroradiometric Measurements,
from the Illuminating Engineering Society (IES 1983). A second ASTM
standard of interest is E 1336, "Standard Test Method for Obtaining
Colorimetric Data from a Video Display Unit by Spectroradiometry."
Also referenced in that standard is a technique prescribed by the
International Electrochemical Commission (IEC), Photometric and
Colorimetric Methods of Measurement of the Light Emitted by a
Cathode-Ray Tube Screen (IEC 1974).
250 Chapter 9
Because radiometry and spectroradiometry are of such great
potential value and importance to those interested in displaying museum
objects and in characterizing illumination in museum environments, the
above standard methods for their use have been listed. Spectroradiometry
and abridged spectroradiometry are far more complex tasks than the
spectrophotometry discussed in these pages; they involve many more
variables, which must be defined and standardized, such as solid angles,
distances, calibrations, and resolutions, in addition to the spectral or col-
orimetric characteristics.
With modern computers, so-called machine color vision is
fast becoming a reality (Novak and Shafer 1992). However, it must be
remembered that progress must be based on a fundamental understand-
ing of spectrophotometry and colorimetry and on the factors affecting
the appearance of materials, as described in this monograph.
No mention has been made in these pages regarding the study
and measurement of retroreflective materials. Retroreflection is defined as
"reflection in which radiation is returned in directions close to the direc-
tion from which it came, this property being maintained over wide varia-
tions of the direction of the incident radiation" (ASTM Standard E 284).
Retroreflective materials are commonly used for traffic safety markings,
license plates, taillights, and so on. Although they may not have been
used by artists as yet, thin films of retroreflective material in sheets and
tapes are readily available in many colors. For details and information
concerning the measurement of retroreflection, consult ASTM Standards
E 808, E 809, E 810, and E 811. CIE Publication No. 54 (1982),
Retroreflection: De-finition and Measurement, presents the international
standard methods . The measurement of retroreflection incorporates the
principles of radiometry and goniophotometry. Special geometric coordi-
nate systems are required and are described in these standards.
Other Instrument Features
The wavelength range of the instrument may offer more versatility in
application for analytical purposes. Some spectrophotometers are
equipped to be used in the ultraviolet or in the near-infrared regions, for
example, in addition to the visible region. Such instruments are generally
designed to measure both reflectance and transmittance factors .
The cost of the instrument is always a consideration when
making a purchase. As a general rule, the most precise instruments are
the most expensive, and t he most versatile instruments are also the most
expensive. The initial cost may not be the most important factor, how-
ever. The importance of the operator and interpreter, and the expense of
suitable training and experience for the persons responsible for the
instrument's proper usage, ultimately may well be the major cost factor.
Instrumentation Overview: The Tasks Determine the Selection
Reports
Regardless of the instrumentation used, all reports should include
sufficient information about the methods of the instrument's use as
well as a precise description-its type, model, serial number, and so
on. ASTM Standard E 805, "Standard Practice for Identification of
Instrumental Methods of Color or Color-Difference Measurement of
Materials," presents an excellent outline of the factors that should be
included in a report.
251
Chapter 10
Suggested Protocol for Recording Spectral
Examination Results
The basic nondestructive techniques to be discussed in this chapter-
that is, those methods of ana lysis that do not require samples to be
taken from an object- are primarily those based on wavelength-
dispersive spectroscopy. They include X-ray fluorescence, which detects
most elements, and ultraviolet, visible (the subject primarily dealt with
here), and infrared electronic absorptions, which indicate chemical
structure. Because many different constituents can exist in any design
or depiction of an object, many areas must be examined to achieve
insight into the materials that might have been used in its creation.
Overlaps of information are common; hence, a systematic scheme of
analysis is essential.
It can be helpful, therefore, to prepare an organized proce-
dure (a protocol) to be followed in approaching the analytical problems,
so that one is not buried in a random collection of measurements made
on many different areas of an object. Fortunately, the number of compo-
nents used in making any single object is often limited, so that by putting
together the information obtained from many measurements, a sense of
the artist's medium and palette can be pieced together.
A recommended approach for organizing visible spectropho-
tometric reflectance data is presented in the following pages. Table 10.1
is an example of a data sheet that could be used in the accumulation of
information from spectral reflectance curve measurements on an object.
The following outline provides explanation and suggestions
for the item listings on the data sheet. It is keyed to their superscript let-
ters or numbers.
Identification of Object
Object Identification: Use some type of coding that uniquely
identifies the object (and its owner).
Artist or Source: If the work is signed, include the name of
the artist (if questionable, use a question mark following the
name); if the ori gi n of the object is known, record it (if ques-
tionable, use a question mark following this designation).
Date of object preparation or creation: If known, record as
such; if an est imate, record as such, wi th a range if possible,
Object identification: Artist or source: Date:
Area measured
Color No. abs. A major A other abs. Positive Negative
Comments
H, V, C bands band bands identification identification
I. Primary colors: B, R, Y
II. Secondary colors: G, P, 0
Ill. Near-neutral colors: Gr, Be, nW
() V>
P>
c -0
"
<
ro
P>

B:

v;
i
O'"
'<
ib c
() V>
.+ ro
P>
"
S,
()
ro
Conclusions
s
0
P> ()
.+
0
P>
-
-
V>
0 ::r
ro
0

.2:
S, ro
P>
()
-0
0
9 0
;I
O'"

....
0
:...
"'
c
(IQ
(IQ
a
..
c..
.,,
0
S-
n
0
O'
"'
..
n
0
a.
:J
(IQ
"'
.,,
..
::i.

m
x
..
3
:J

0
:J
"'
:J:
c

"'
\11
UJ
254 Chapter 10
for example, 1940-50, again using a question mark if this
range is uncertain.
Working Guide Notes
Area Measured: Assign a number to the measured area and, if
possible, identify it by a tracing on the back of the object or
on a photocopy of the object or of a picture of the object,
and include a description of the area, such as the sky or grass
of a landscape painting, or the clothing, drapery, or facial
colors of a portrait. Possibly an x and y dimension can be
used based on the distance from the lower left corner, as in
the usual two-dimensional graph. For a three-dimensional
object, provide a location as best as possible.
Color: Include the color name of the hue (H) being measured,
for example, the blue (B) in the sky or water, plus an esti-
mate of the lightness or darkness of the measured area (the
Munsell Value, V), using the abbreviations of L for light (V =
6-9), M for mid-value (V = 3-5), and D for dark (V = 0-2).
To complete the color description, estimate the chroma or
saturation: HC for high-chroma colors, MC for medium-
chroma colors, and LC for low-chroma colors (near neutrals).
Thus, a typical description of the measured color of a painted
sky could be B, M, LC, indicating a blue (B) of moderate
lightness (M) and low chroma (LC). For additional modifiers,
see the remarks that follow the outline.
Number of Absorption Bands: This is important to note; it
indicates that pigments exhibiting more than one absorption
band may be present or that there may be other colorants
present.
A, of Major Absorption Band: The symbol A. is used for wave-
length in nanometers (mi llimicrons in older literature) .
A, of Other Absorption Band(s}: Present the bands in the
order of intensity and, secondarily, on the basis of wave-
lengt h-long to short.
Positive Identification: If a unique characteristic is observable
in the spectral curves, note it here. For example, phthalo blue
and quinacridone structures are often unmistakable.
Negative Identification: Note pigments that are probably not
present. For example, in a blue region note if phthalo blue
with its characteristic red tail, or iron blue with its dropping
reflectance in the red region, are not present . To avoid confu-
sion, do this for each color area of the work measured. Do
not try to interpret the evidence on the basis of a limited
number of measured curves. Facts established step by step
eventually reveal a total picture.
Suggested Protocol for Recording Spectral Examination Results
Comments: Primarily describe uncertain or questionable
results here. It will help with subsequent interpretations.
255
Conclusions: Include any positive identifications, possible
pigments, and recommendations for further measurements,
analysis, and testing. If the spectral information is to be used
for selecting the least metameric alternate composition for
use in reproducing the color, the most suitable colorant or
colorant combination could be noted here.
It may be desirable to modify the color designations in col-
umn 2 of the data sheet. More specific relative descriptions of the hue,
value, and chroma estimates of column 2 may be provided by the use of
simple modifiers. Typical hue designations can be B for blue, R for red,
Y for Yellow, G for green, P for purple (or violet), and 0 for orange.
For the near neutrals, Gr is used for gray, Be is used for beige, and n W
is used for near white. It is a matter of judgment as to when a green
becomes a blue or a yellow becomes an orange, and so on. The hue
designations selected by one examiner may vary somewhat from those
selected by another. Using additional hue modifiers may be of help in
avoiding large discrepancies between individuals: differentiations of the
basic hue selected may be described by adding a lower-case prefix. For
example, measurements made on a blue sky in different areas may be
described as rB to indicate a blue area that is redder than the basic blue,
or gB to describe a blue area that is greener. Similar descriptors can be
used for all of the nine basic hue categories.
Likewise, the relative chroma or lightness can be modified by
use of a plus sign ( +) or a minus sign (-) after the chroma or lightness
designation. Thus, for subsequent measurements of the blue sky illus-
trated above, a designation could be rB, M+, LC- to describe a blue that
is redder than the area designated B; is lighter; and is lower in chroma.
The above descriptors are based on visual evaluations for the
convenience of the examiner and are, therefore, relative. They are not
defined in any color-order system per se but could be based on reference
to the Munsell Book of Color if desired.
The data sheet illustrated is abbreviated in length-more data
would undoubtedly be needed. However, the important aspect of this
abbreviated chart is the systematic approach. A similar but suitably
modified protocol could be followed for use with X-ray fluorescence
analyses for the metals.
After making many detailed inspections, measurements,
and evaluations of the various areas, one should then try to put it all
together: enumerate the pigments found with relative certainty; make a
record of those not found; note those that may be questionable and sub-
ject to further identification techniques if possible; and prepare a sum-
mary of the probable palette (selection of pigments) used in creating the
object. Further tests that may be required should be noted.
Other regions of absorption are the ultraviolet and the near
infrared. Little has been published about pigment absorptions in the
ultraviolet. It is probable, however, that only reflectances measured in
256 Chapter 10
the near ultraviolet (above 300 nm) would be of value in the case of
paints, because of vehicle absorption at shorter wavelengths. The near-
infrared region, 700-1000 nm, has been studied more extensively,
primarily for camouflage applications. The Colors Division of Ciba
Specialty Chemicals (formerly Ciba-Geigy) supplies spectral curves in
this region as well as in the visible. For certain pigments, reflectances
measured in the near-infrared region can be helpful in identification. The
reflectance curves for the a- and blues, for example,
differ markedl y in this region. Many of the synthetic inorganic pigments
tailored for use in camouflage materials likewise have characteristic spec-
tra in the infrared. Whether further study of the regions adjacent to the
visible would be helpful is not fully known. Perhaps the near-ultraviolet
region would be worthy of further study, especially because of the actinic
effect of this shorter-wavelength region.
Finall y, it must again be emphasized that in any analytical
task, reference material based on known compounds is required.
Reference reflectance curves in the visible spectrum of colored com-
pounds are no exception.
Chapter 11
Summary, Conclusions, and Recommendations
The interpretation of spectrophotometric reflectance and transmittance
curves in the visible wavelengths encompasses many aspects of color sci-
ence in addition to the identification of colorants. It is hoped that the
material presented in this monograph will prove useful in solving a num-
ber of challenging problems encountered in the care of museum objects.
The ability to properly interpret spectrophotometric data (the curves)
will be a valuable aid in identifying the colorants used in the original
work so that the best care and methods of preservation can be instituted.
It will contribute to the selection of the colorants most suitable for
reproducing the color when necessary. It will provide for the detection
and avoidance of metamerism. It will aid in determining when analytical
terms of the measured data should be used, or when transformations to
trichromatic terms would be more appropriate. It will aid in determining
the most meaningful measurement technique for particular materials,
such as metals, and for various conditions, such as mottling, streaking,
or other irregularities, in the areas to be measured. It will help in deter-
mining types and rates of deterioration, and in describing them in analyt-
ical terms that can be used for objective interpretation. It will aid in
calculating colorant mixture formulations to make desired colors (after
suitable reference curves of known pigment concentration are prepared).
At the very least, the ability to interpret spectrophotometric data prop-
erly contributes to a language that can be used to describe one of our
sensations in universally accepted terminology. These points represent the
author's principal objectives in preparing this manuscript.
The author's orientation toward the use of pigments led to
discussions of the many newer types of pigments and their measure-
ment-fluorescent pigments, pearlescent and interference pigments, and
microvoids and vesiculated beads, as well as the more traditional metallic
flake pigments. Although these pigments may not occur in antiquities,
their presence may occur in post-World War II artifacts or works of art.
Most of the examples presented involve pigmented systems; however, the
same basic principles apply to materials colored with dyes, such as tex-
tiles, paper, and some plastics.
It is hoped that the procedures and guidelines outlined will
help the user avoid inappropriate color measurements-measurements
made with instruments not carefully standardized; measurements made
with instruments not suitable for the task at hand; measurements made
258 Chapter 11
with instruments whose reliability and reproducibility are not known;
or measurements from which improper conclusions can be drawn out
of ignorance. As has been said many times about many specialized ana-
lytical applications of instrumentation, the knowledgeable analyst can
get a great deal of valid information from the proper use of even the
most simple equipment. Conversely, those not experienced in the evalu-
ation of data from elaborate, sophisticated equipment can accumulate
considerable information that is meaningless and interpret it with
unjustified conclusions.
One cannot understand color and the problems associated
with its characterization and description without appreciating its inter-
disciplinary nature. Color is a psychological response to a physical stimu-
lus. Moreover, object color is but one aspect of appearance. Appearance
is affected by many factors: surface characteristics (glossiness, matteness,
and uniformity), internal composition, homogeneity, size, shape, and
contrast with the surroundings. The attempt to measure and describe
human perception requires the application not only of the physical
sciences (physics, chemistry, mathematics) but also of the psychology
and physiology of vision. Thus, color science requires some knowledge
of light and its measurement; some knowledge of the chemistry of the
materials that contribute to color; some knowledge of mathematics
for treating the measured data; and some knowledge of the visual mecha-
nism and factors that affect the brain's processing of the sensations into
a perception.
Because of its interdisciplinary nature, the science of color
per se is taught at only a few educational institutions. It is a specialty
learned primarily by studying the contributions of previous experts;
many of them acquired much of their understanding on their own initia-
tive, helping to create the body of knowledge we have today. The books
and selected references listed in appendix E, most of which have already
been referred to in the text, provide a useful introduction as well as a
means for continuing study. Short courses of a few days to a week, often
presented by vendors of instruments, can be helpful if it is remembered
that the vendors' primary purpose may be to sell something. A few uni-
versities do present short courses, some of an advanced nature and some
for a specialized area of application. Nonetheless, the successful acquisi-
tion of knowledge about color science depends ultimately on personal
pursuit and study.
From the material on spectrophotometry and colorimetry pre-
sented in the text, it should be obvious that a reference library of spec-
trophotometric curves is essential to the application of many of the
techniques described. Instead of each laboratory collecting and preparing
its own atlas of spectral curves, the information already available in pub-
lications or in other laboratories should be collected together for the use
of everyone. The development of color-measuring instruments and com-
puter color-matching techniques in the 1950s led to the measurement and
publication of numerous reflectance curves, particularly by colorant man-
ufacturers. Before many of the curves become lost due to obsolescence,
Summary, Conclusions, and Recommendations 259
changes in manufacture, or lack of interest, they should be accumulated
into an atlas.
Collections of dyes as well as pigments of known origin and
composition should also be accumulated and updated as new materials
become available. It would be a shame to lose such a valuable record of
the newer colorants that are being used by artists and artisans today-
they will be of great value for conservation scientists in the future.
With the expanding use of color-measuring instruments in
museum laboratories, collections of both colorants and colorant spec-
trophotometric curves would aid tremendously in the study of the vari-
ous problems faced. This is particularly true for the care of relatively
modern works, some of which are already in need of restoration and all
of which need proper environmental conditions for their preservation.
Organizing the hundreds of curves in a manner that facili-
tates their use presents a challenge. A protocol for one type of orderly
arrangement is presented in appendix D; it is based on hue as well as on
the number, location (wavelength), and relative intensity of the absorp-
tion bands exhibited in the spectral reflectance curves. The organization
of some of the many red pigments available is used as an example.
Preparation of such a library may appear initially to be a formidable
task, particularly if the curves were to be redrawn to a standard scale.
However, if an institution or organization were to take on the work of
accumulating and organizing the curves and of offering them simply as
reproductions in their original format, it would be carrying out an
invaluable service. The author would be happy to turn over her collec-
tion as a start in this effort. Many companies and laboratories would be
willing to contribute copies of their spectral curves for this undertaking.
It is the author's sincere hope that such an atlas of spectrophotometric
curves will be made in the near future.
Appendix A
Curves and Data for Pigments Used as
Illustrative Problems
The reflectance curves, CIE colorimetric data, and Munsell notations for
the colorants used as illustrations in the text are presented for eleven
chromatic colorants, four white pigments, and a carbon black at a series
of concentrations. All of the samples were prepared in a high gloss ther-
mosetting acrylic automotive enamel. The curves for the eleven chromatic
colorants and the carbon black are mixtures with rutile Ti02 white;
curves for one concentration of the eleven chromatic colorants and rutile
Ti02 with carbon black as well as with burnt umber are also included.
For the transparent pigments, only the maximum concentrations used in
white are illustrated; for scattering pigments, masstones are also illus-
trated. Masstone curves for the four whites as well as mixtures with car-
bon black are included. Curves for the whites, black, and burnt umber
are presented first. Then the curves for the primary colors, blue, red, and
yellow are illustrated, followed by the curves for the secondary colors,
green, violet (purple), and orange. Because the secondary colors and the
near-neutral colors can be made by admixture of appropriate primary
colorants, a few curves of such mixtures are also included.
The reflectance curves illustrated are the total reflectances-
that is, the surface specular reflections are included. The CIE colorimetric
data for illuminant C and the 2 observer were determined for the diffuse
reflectances obtained by subtracting 4 % for the specular surface
reflections from the total reflectances. The CIE data are illustrated on a
chromaticity diagram in figure A. l. For those samples whose masstones
are included, they are connected to the highest concentration in white by
a dotted line.
The CIE colorimetric data and the Munsell notations calcu-
lated from these values are presented in tables that are interspersed with
the reflectance curves according to hue.
For simplicity in presenting the CIE and Munsell notations in
the tables, abbreviations for the pigments have been used. The following
is a list of them:
R. Ti02
A. Ti02
ZnO
Pb white
Black
rutile titanium dioxide; PW 6
anatase titanium dioxide; PW 6
zinc oxide; PW 4
basic lead carbonate; PW 1
carbon black; Pbk 7
Curves and Data for Pigments Used as Illustrative Problems 261
UB
PB
ROX
BON
QR
YOX
FGL
PG
CV
MO
ultramarine blue; PB 29
alpha-phthalocyanine blue (phthalo blue, red
shade); a-PB 15
synthetic red oxide; PR 101
BON red (BON is the abbreviation for beta-
oxynaphthoic acid coupling component used
in the making of many red pigments. Many
red pigments were formerly called BON reds,
so the exact composition is not known.)
Quinacridone Red
synthetic yellow iron oxide; PY 42
Permanent Yellow FGL; PY 79
phthalocyanine green
carbazole dioxazine violet; PV 23
molybdate orange; PR 104
Colorant names are often misleading; the same name may be
applied to different colorants. An example is the BON red described
above. Accordingly, the need for a method that would identify colorant
chemical composition was recognized long ago. The system most widely
used today to accomplish this purpose is the Colour Index published
under the aegis of the Society of Dyers and Colourists of Great Britain
and the American Association of Textile Chemists and Colorists. It is
described by Billmeyer and Saltzman. A brief description of the system is
given here in appendix C.
In most application laboratories, Colour Index names and
numbers were not used extensively when the samples for the reflectance
curves in this appendix and those in Chapter 7 were made. At that time,
popular color names or trade names were employed in everyday use.
Whenever possible, the Colour Index names, such as PW 6 for the Ti02
white pigments, have been used in the abbreviation list above. Those
appearing without Colour Index names cannot be cited with certainty at
this late date. This omission is generally not serious, because a Colour
Index name alone is not generally sufficiently specific. For example, PW
6 does not distinguish between rutile and anatase Ti02. The Colour
Index number is more specific but also does not characterize all of the
variations in pigments that may be dependent on variations in manufac-
turing and treatments.
262
Figure A.1
Chromaticity diagram of colorants in
appendix A mixed with rutile Ti0
2
.
Appendix A
0.9
520
0.8
0.7
0.6
::>.. 0.5
0.4
0.3
0.2
0.1
5. ultramarine blue
6. phthalocyanine blue
7. red iron oxide
8. BON red
9. Quinacridone Red
10 yellow iron oxide
11. Permanent Yellow FGL
12. phthalocyanine green
13. carbazole dioxazine violet
14. molybdate orange
@ masstone
@ max. cone.
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
x
Figure A.2
Portion of the chromaticity diagram
for masstone whites and white-black
mixtures.
Figure Pigment composition
A.3 0.10% black, 99.9% R. Ti0
2
0.50% black, 99.5% R. Ti0
2
2.00% black, 98.0% R. Ti0
2
8.00% black, 92.0% R. Ti0
2
A.4 1.00% black, 99.0% R. Ti0
2
0.95% black, 99.05% A. Ti0
2
0.19% black, 99.81 % ZnO
0.23% black, 99.77% Pb whit e
A.5 100% R. Ti0
2
100% A. Ti0
2
100% ZnO
100% Pb white
A.6 0.5% B. umber, 99.5% R. Ti0
2
2.5% B. umber, 97.5% R. Ti0
2
10.0% B. umber, 90.0% R. Ti0
2
40.0% B. umber, 60.0% R. Ti0
2
100.0% B. umber, 0.0% R. Ti0
2
Table A.1
CIE and Munsell notation (diffuse
reflectance) for whites, blacks, and
burnt umber.
Curves and Data for Pigments Used as Illustrative Problems
0.32
0.31
0.30
0.29
white pigment: masstones
and mixtures with black
R.Ti02
0 Pb white
/0.23%Bk
1.0%Bk
A.Ti02 I
Oo95%Bk
b,. ZnO /
0.19%Bk
263
0.28 - + - - - - - - ~ - - - - ~ - - - ~ - - - - ~ - - - ~
0.28 0.29 0.30 0.31 0.32 0.33
x
CIE Munsell
x y
y
Hue Value/Chroma
0.3043 0.3120 68.32 1 .61 PB 8.50/ 0.79
0.2982 0.3061 48.45 3.1 4 PB 7.36/1 .54
0.2909 0.2990 29.06 3.27 PB 5.92/ 1.93
0.2830 0.2907 13.21 3.08 PB 4.18/ 1.83
0.2947 0.3030 38.65 3.20 PB 6.69/ 1.78
0.2881 0.2952 36.98 4.08 PB 6.56/ 2.55
0.2853 0.2902 39.57 5.32 PB 6.76/ 3.08
0.3016 0.3082 41 .34 4.49 PB 6.89/ 1.09
0.3129 0.3203 90.66 8.41 y 9.53/ 0.22
0.3103 0.3165 91.38 0.00 N 9.56/ 0.00
0.3180 0.3257 86.30 4.37 y 9.35/ 0.57
0.3178 0.3241 86.16 1.58 y 9.34/ 0.53
0.3266 0.3286 71.21 7.17 YR 8.65/ 1. 19
0.3416 0.3377 52.38 6.63 YR 7.61/ 2.08
0.3620 0.3487 32.30 6.30 YR 6.1 9/ 2.83
0.3905 0.3594 12.82 6.07 YR 4.13/ 3.06
0.3790 0.3384 1.81 4.09 YR 1.39/1. 38
264 Appendix A
100
90
80
70
"irj 60
g
"'
u
c
.s
u
50
"'
q::

..._,
c
"'
u
a;
40 CL
30
10
Figure A.3 400
Carbon black in rutile Ti0
2
.
0.0
0.1 %
450
y 0
black mixtures in rutile Ti02
500 550
Wavelength (nm)
600
-----R----
650 700
Figure A.4
Carbon black in rutile Ti0
2
, anatase Ti0
2
,
ZnO, and Pb white.
Curves and Data for Pigments Used as Illustrative Problems
100
90
bO
70
60
50
40
30
20
10
400 450
G
y
black mixtures in whites
1.0% black in rutile Ti0
2
0.95% black in anatase Ti0
2
0.1927% black in ZnO
0.2275% black in Pb white
0
500 550 600
Wavelength (nm)
265
-----R----
650 700
266
Figure A.5
Masstone whites: rutile Ti0
2
, anatase Ti0
2
,
ZnO, and Pb white.
Appendix A
G
y
0 i-.---R----
100
.
...

90
..,.... ..... ..,,.... .. = ~ -=-- :-:
: ---
----
80
,
I
70
I
60
50
40
30
20
10
400
-----
---
. ...r
;
,
450
masstone whites
--- rutile Ti02
anatase Ti0
2
ZnO
--Pb white
500 550 600
Wavelength (nm)
650 700
Figure A.6
Burnt umber mixtures in rutile Ti0
2
.
Curves and Data for Pigments Used as Illustrative Problems 267
100
90
80
70
"El
60
g
Q)
u
c:
<tj
t
50 Q)
;:;::

.....
c:
Q)

Q)
40 CL
30
20
10
400 450
G
y
0 -----R----
burnt umber mixtures in rutile Ti02
40.0%
100.0%
500 550
Wavelength (nm)
600 650 700
268 Appendix A
0.39 ....------------------------------.
Figure A.7
Portion of the chromaticity diagram for the
chromatic colorants mixed with rutile Ti0
2
and with black or burnt umber added.
Figure Pigment composition
A.8 3.00% UB, 97.0% R. Ti0
2
12.50% UB, 87.5% R. Ti0
2
50.0% UB, 50.0% R. Ti0
2
75.0% UB, 25.0% R. Ti0
2
100.0% UB, 0.0% R. Ti0
2
0.37
0.35
0.33
0.31
0.29
0.27
A.9 12.5% UB, 0.02% black, 87.48% R. Ti0
2
12.5% UB, 0.10% black, 87.40% R. Ti0
2
12.5% UB, 0.5% B. umber, 87.00% R. Ti0
2
A.10 0.10% PB, 99.9% R. Ti0
2
0.25% PB, 99.75% R. Ti0
2
0.50% PB, 99.50% R. Ti0
2
1.00% PB, 99.00% R. Ti0
2
2.00% PB, 98.00% R. Ti0
2
A.11 0.25% PB, 0.10% black, 99.65% R.Ti0
2
0.25% PB, 0.50% B. umber, 99.25% R. Ti0
2
Table A.2
CIE and Munsell notation (diffuse reflec-
tance) for ultramarine blue (U B) and
phthalo blue (PB).
chromatic colorants in rutile
Ti0
2
with black or b. umber
0.26 0.28 0.30
CIE
x y
0.2896 0.2990
0.2694 0.2779
0.2313 0.2324
0.2058 0.1951
0.1475 0.0645
0.2711 0.2800
0.2750 0.2845
0.2864 0.2983
0.2703 0.2900
0.2526 0.2747
0.2370 0.2602
0.2201 0.2432
0.2029 0.2242
0.2614 0.2827
0.2712 0.2965
0.32
x
y
71.42
54.35
28.56
16.63
2.27
53.33
49.97
50.10
65.26
54.84
46.14
37.27
28.77
50.19
49.96
0
/ / / yellow
YOXO
~ ~
10.0
/ FGL 1.0%
0.34
~
O no black, no b. umber
,6. plus 0.10% black
plus 0.50% b. umber
0.36 0.38 0.40
Munsell
Hue Value/Chroma
2.46 PB 8.66/ 2.66
3.82 PB 7.73/ 4.85
4.79 PB 5.87 / 8.33
5.57 PB 4.64/10.20
6.28 PB 1.63/17.45
3.75 PB 7.67/ 4.63
3.58 PB 7.46/ 4.10
1.33 PB 7.47/ 2.68
0. 14 PB 8.34/ 4. 13
0.60 PB 7.76/ 5.89
0 .73 PB 7.21 /7.33
0.94 PB 6.59/ 8.72
1.24 PB 5.89/ 10.06
0.19 PB 7.47 / 4.87
7.45 B 7.46/ 3.62
Curves and Data for Pigments Used as Illustrative Problems 269
G
y
0 i.---R----
100
ultramarine blue mixtures in rutile Ti0
2
90
80
70
3.0%

60
g
"'
u
c:
_:g 12.5%
u
50
"'
;:;:::

....,
c:
"'
"'
40 Q_
30
50.0%
20
75.0%
10
100.0%
Figure A.8 400 450 500 550 600 650 700
Ultramarine blue mixtures in rutile Ti0
2
. Wavelength (nm)
270
Figure A.9
Ultramarine blue mixtures in rutile Ti0
2
,
and with carbon black or burnt umber
added.
Appendix A
v i ~
B _ __.,14-__
G y 0 !-.--- R ----
90
80
70
2
60
g
v
u
c
cu
t)
50
v
""' l':'
.._,
c
v
~
v
40
0...
30
20
10
400
ultramarine blue mixtures in rutile Ti0
2
plus carbon black or burnt umber
-- 12.5% blue+ 0.1 % black
12.5% blue+ 0.5% burnt umber
- - 12.5% blue+ 0.02% black
-- 12.5% blue+ 87.5% Ti02
450 500 550
Wavelength (nm)
600 650 700
Figure A.10
Phthalocyanine blue mixtures in rutile Ti0
2
.
Curves and Data for Pigments Used as Illustrative Problems
90
80
70
<il 60
0

"'
u
c
JS
u
50
"'
q::

....,
c
"'
u
cu
40 a_
30
20
10
400 450 500
y 0
phthalocyanine blue
mixtures in rutile Ti0
2
550
Wavelength (nm)
600
271
R----
650 700
272
Figure A.11
Phthalocyanine blue mixtures in rutile
Ti0
2
and with carbon black or burnt
umber added.
Appendix A
--v --'""*4---
G
y
0
100
90
80
I
70
l'i
60
g
Q)
u
c
_::g
u
50 Q)
;:;:::

.....
c
Q)
u
Q;
40
"-
30
20
10
400
phthalocyanine blue mixtures in rutile Ti0
2
plus carbon black or burnt umber
,,.,-- ... ,,
, '
, \
, \
I \
...
\
\
0.25 % blue + 0.1 % black
450
0.25 % blue+ 0.5% burnt umber
0.25% blue
500 550
Wavelength (nm)
600
R----
650 700
Curves and Data for Pigments Used as Illustrative Problems
Figure Pigment composition
A.12 0.50% ROX, 99.5% R. Ti0
2
2.00% ROX, 98.00% R. Ti0
2
12.5 % ROX, 87.5% R. Ti0
2
100.00% ROX, 0.00% R. Ti0
2
A.13 2.00% ROX, 0.10% black, 97.9% R. Ti0
2
2.00% ROX, 0.50% B. umber, 97.5% R. Ti0
2
A.14 0.25 % BON, 99.75 % R. Ti0
2
0.50% BON, 99.5 % R. Ti0
2
1.00% BON, 99.00% R. Ti0
2
4.00% BON, 96.00% R. Ti0
2
15.00% BON, 85.00% R. Ti0
2
100.00% BON (masstone)
A.15 0.50% BON, 0.10% black, 99.4% R. Ti0
2
0.50% BON, 0.50% B. umber, 99.00% R. Ti0
2
A.16 0.50% QR, 99.5% R. Ti0
2
1.00% QR, 99.00% R. Ti0
2
2.00% QR, 98.00% R. Ti0
2
4.00% QR, 96.00% R. Ti0
2
10.00% QR, 90.00% R. Ti0
2
A.17 0.50% QR, 0.10% black, 99.40% R. Ti0
2
0.50% QR, 0.50% B. umber, 99.00% R. Ti0
2
Table A.3
CIE and Munsell notation (diffuse
reflectance) for red iron oxide (ROX),
BON red, and Quinacridone Red (QR).
CIE
x y
y
0.3374 0.3209 67.31
0.3595 0.3235 50.89
0.4130 0.3304 26.76
0.5585 0.3595 7.85
0.3476 0.3215 47.00
0.3595 0.3261 48.47
0.3320 0.3062 68.80
0.3409 0.3028 61 .61
0.3530 0.2990 53.52
0.3918 0.2922 36.36
0.4557 0.2910 21.91
0.6157 0.3248 7.32
0.3225 0.3018 54.44
0.3407 0.3113 56.60
0.3340 0.2954 63.93
0.3431 0.2890 56.11
0.3553 0.2882 47.65
0.3715 0.2756 39.00
0.4012 0.2685 28.11
0.3172 0.2960 56.01
0.3366 0.3080 58.29
273
Munsell
Hue Value/Chroma
4.88 R 8.45/ 3.05
5.27 R 7.52/ 4.49
6.20 R 5.71/6.80
0.66 YR 3.28/ 9.40
5.05 R 7.27/ 3.47
6.42 R 7.36/ 4.19
5.59 RP 8.53/4.06
6.00 RP 8.14/ 5.13
6.31 RP 7.68/ 6.46
8.00 RP 6.52/ 9.62
0.58 R 5.24/1 1.89
8.42 R 3.17/1 2.28
2.58 RP 7.73/ 3.52
8.92 RP 7.86/ 4.14
3.34 RP 8.27/ 5.48
3.58 RP 7.83/7.04
4.1 2 RP 7.31 / 8.41
4.73 RP 6.71 / 10.1 6
6.05 RP 5.83/ 12.00
9.99 p 7.83/ 3.83
7.08 RP 7.96/ 4.17
274 Appendix A
90
80
70
<ii 60
0
:::.
Q)
u
c
_:g
u
50 Q)
;+:::

...,
c
Q)

Q)
40 Q._
30
20
10
Figure A.12
400
Red iron oxide mixtures in rutile Ti0
2
.
450
red iron oxide mixtures in rutile Ti0
2
0.5 %
20%
12.5%
100.0 %
500 550
Wavelength (nm)
600 650 700
Figure A.13
Red iron oxide mixtures in rutile Ti0
2
, and
with carbon black or burnt umber added.
Curves and Data for Pigments Used as Illustrative Problems 275
100
90
80
70
60
50
40
30
20
10
400
G
y
0
.__---R----
------
red iron oxide mixtures in rutile Ti0
2
plus carbon black or burnt umber
....
.. ..
; ....... .
, ......
, ......... .
I ,.'"
I ,.'"
1:
1:
1:
1:
1:
1:
1:
1:
----
....................... ::----
.
450
2.0% red+ 0.1 % black
2.0% red+ 0.5 % burnt umber
2.0% red
500 550
Wavelength (nm)
600 650
,
,
. ..
. ..
700
276 Appendix A
-V ---i-11-- B -------
G y 0
100 .--.--.--.--t--....--i....--i....--i----.--;----.---.---.---r---r---r--+---r--+---r--+---r---.---.---.---.---.---.---.---.
BON red mixtures in rutile Ti0
2
90
80
70
]i
60
g
"'
u
c
.::;l
u
50
"'
;;:::

....,
c
"'
"'
40
CL
30
20
10
100.0%
Figure A.14 400 450 500 550 600 650 700
BON red mixtures in rutile Ti0
2
. Wavelength (nm)
Figure A.15
BON red mixtures in rutile Ti0
2
, and with
carbon black or burnt umber added.
Curves and Data for Pigments Used as Illustrative Problems 277
100
90
80
70
<il 60
g
<lJ
u
c
(lj
....,
u
50
<lJ
q::

....,
c
<lJ
u
Q:;
40 CL
30
20
10
400 450
G
y
0 ------R----
BON red mixtures in rutile Ti0
2
plus carbon black or burnt umber
./
I
.
,,.....----
I
.
I
...
. ..
I::
..
//
..
/ .. :
., L'.:'
.,_ :'
'----1
0.5 % red+ 0.1 % black
0.5 % red+ 0.5 % burnt umber
0.5 % red+ 99.5 % Ti0
2
500 550 600
Wavelength (nm)
650 700
278
Figure A.16
Quinacridone Red mixtures in rutile Ti0
2
.
Appendix A
-v--:i.i..-- B ~ 1 4
G
90
80
70
Iii
....,
60
g
<!)
u
c
.s
u
50
<!)
q::
l.':'
....,
c
<!)
~
<!)
40
0..
30
20
10
400
Quinacridone Red mixtures in rutile Ti0
2
450 500 550
Wavelength (nm)
600 650 700
Figure A.17
Quinacridone Red mixtures in rutile Ti0
2
,
and with carbon black or burnt umber
added.
Curves and Data for Pigments Used as Illustrative Problems
100
90
G
y
Quinacridone Red mixtures in rutile Ti0
2
plus carbon black or burnt umber
I
I
0
I
;
,
279
R----
----------
80

I .
I .
70
60
50
40
30
20
10
400
-...
I '
I '
'
'
'
.
I ,
I .'
I .'
I.'
1:
1:
1:
1:
1:
I
--- 0.5% red+ 0.1 % black
450
0.5% red+ 0.5% burnt umber
0.5% red
500 550
Wavelength (nm)
600 650 700
280 Appendix A
Figure Pigment composition
A.18 1.00% YOX, 99.00% R. Ti0
2
2.50% YOX, 97.50% R. Ti0
2
10.00% YOX, 90.00% R. Ti0
2
25.00% YOX, 75.00% R. Ti0
2
100.00% YOX (masstone)
A.19 10.00% YOX, 0.10% black, 89.9% R. Ti0
2
10.00% YOX, 0.5% B. umber, 89.5% R. Ti0
2
70.00% YOX, 5.00% black, 25.00% R. Ti0
2
A.20 0.50% FGL, 99.5% R. Ti0
2
1.00% FGL, 99.0% R. Ti0
2
2.50% FGL, 97.5% R. Ti0
2
5.00% FGL, 95.00% R. Ti0
2
10.00% FGL, 90.00% R. Ti0
2
A.21 1.00% FGL, 0.10% black, 98.9% R. Ti0
2
1.00% FGL, 0.5% B. umber, 98.5% R. Ti0
2
10.00% FGL, 1.00% black, 89.0% R. Ti0
2
Table A.4
CIE and Munsell notation (diffuse
reflectance) for yellow oxide (YOX)
and Permanent Yellow FGL.
90
80
70
<il 60
...,,
g
cu
u
c
l'l
u
50
cu
;:;:::

...,,
c
cu

cu
40
"-
30
20
10
Figure A.18 400
Yellow iron oxide mixtures in rutile Ti0
2
.
CIE Munsell
x y
y
Hue Value/Chroma
0.3411 0.3453 80.07 0.41 y 9.07/2.14
0.3565 0.3577 73.63 0.15 y 8.77/3.19
0.3909 0.3823 59.68 9.82 YR 8.04/5.43
0.4234 0.4017 47.67 9.46 YR 7.31/7.05
0.4849 0.4387 24.18 9.86 YR 5.47 /9.27
0.3756 0.3772 53.00 1.51 y 7.65/4.27
0.3874 0.3793 55.31 9.83 YR 7.79/5.06
0.3732 0.4044 10.88 9.09 y 3.83/3.06
0.3471 0.3642 88.21 0.13 GY 9.43/2.67
0.3531 0.3805 87.15 9.71 y 9.38/3.72
0.3761 0.4066 85.21 8.91 y 9.30/5.49
0.3904 0.4285 83.23 8.25 y 9.21/7.08
0.4100 0.4503 80.64 7.66 y 9.10/8.94
0.3317 0.3583 66.79 4.24 GY 8.42/2.25
0.3493 0.3637 69.99 4.79 y 8.59/2.87
0.3530 0.4033 35.67 3.69 GY 6.46/4.26
y 0 -----R----
yellow iron oxide mixtures in rutile Ti0
2
100.0%
450 500 550
Wavelength (nm)
600 650 700
Figure A.19
Yellow iron oxide mixtures in rutile Ti0
2
,
and with carbon black or burnt umber
added.
Curves and Data for Pigments Used as Illustrative Problems 281
100
90
80
70
] 60
g
"'
u
c
"' t)
"'
;:;:::
~
.....
c
"'
"'
Cl.
50
40
30
20
10
400
G
y
0 R----
yellow iron oxide mixtures in rutile Ti0
2
plus carbon black or burnt umber
,,,.---. -
/
. ---
.1 .. ................... .
I .. ~ ~

1 ... :
--- 10.0% yellow+ 0.1 % black
---
---
450
10.0% yellow+ 0.5% burnt umber
70.0% yellow + 5.0% black
10.0% yellow
---
.;' ----- / ----
/
--
500 550
Wavelength (nm)
600 650 700
282 Appendix A
100
90
80
70
}
60
g
"'
u
c
1l
u
50
"'
;;:::

....,
c
"'
"'
40 CL
30
20
10
Figure A.20
400
Permanent Yellow FGL in rutile Ti0
2
.
450 500
y 0
550
Wavelength (nm)
600
----- R----
650 700
Figure A.21
Permanent Yellow FGL in rutile Ti0
2
, and
with carbon black or burnt umber added.
Curves and Data for Pigments Used as Il lustrative Problems 283

G
y
0 R----
Permanent Yellow
90
80
70
ca 60
g
FGL mixtures in rutile Ti0
2
plus carbon black or
burnt umber
/
.
I
.
I
.
I
.
I
.
I
..
..
..
..
..
..
. ....
....................
....

/ ----------
---
30
I
/
/
I
I
I
I
20 - - - ...-
10
400
1.0% yellow + 0.1 % black
1.0% yellow+ 0.5% burnt umber
10.0% yellow + 1.0% black
1.0% yellow
450 500 550
Wavelength (nm)
---
--
600 650 700
284 Appendix A
Figure Pigment composition
A.22 0.10% PG, 99.9 % R. Ti0
2
0.25 % PG, 99.75 % R. Ti0
2
0.50 % PG, 99.50% R. Ti0
2
2.00 % PG, 98.00% R. Ti0
2
4.00 % PG, 96.00 % R. Ti0
2
A.23 0.25 % PG, 0.10% black, 99.65% R. Ti0
2
0.25 % PG, 0 .50% B. umber, 99.25% R. Ti0
2
A.24 0.30 % PB, 3.0 % FGL, 96.7 % R. Ti0
2
0.20 % PB, 5.00 % YOX, 94.8% R. Ti0
2
25.00 % UB, 4.00% FGL, 71 .00% R. Ti0
2
21.00 % UB, 5.00% YOX, 74.00% R. Ti0
2
CIE and Munsell notation (diffuse
reflectance) for phthalocyanine green (PG)
and green mixtures (with yellow oxide
[YOX] and Permanent Yellow FGL).
Figure A.22
Phthalocyanine green mixtures in
rutile Ti0
2
.

g
Q)
u
c
_:g
u
Q)
:;::
[L'
......
c
Q)
u
OJ
Q_
100
90
80
70
60
50
40
30
20
10
400
CIE
x y
0.2877 0 .3216
0.2768 0 .3207
0 .2667 0 .3260
0 .2422 0.3341
0 .2284 0.3402
0 .2825 0.3163
0 .2934 0.3285
0.2996 0 .3744
0.3140 0 .3459
0.3134 0 .3742
0.3104 0.3333
B
450 500
y
77.41
70.28
63 .58
47.49
38.72
60.52
61.02
48.74
47.59
39.51
39.51
y
550
Wavelength (nm)
Munsell
Hue Value/Chroma
6.59 BG 8.95/ 2.49
6.38 BG 8.60/ 3.66
6.13 BG 8.25/ 4.83
4.86 BG 7.30/7.53
4.23 BG 6.69/ 8.67
8.59 BG 8.08/ 2.70
1.75 BG 8.11 / 2.18
2.07 G 7.38/5.04
9 .15 GY 7.31 / 1.86
9 .91 GY 6.75/3 .75
0.41 G 6.75/ 1.12
0 R
phthalocyanine green
mixtures in rutile Ti02
2.0%
4.0%
600 650 700
Figure A.23
Phthalocyanine green mixtures in rutile
Ti0
2
, and with carbon black or burnt
umber added.
Curves and Data for Pigments Used as Illustrative Problems 285
90
80
70
]j 60
g
40
30
20
10
400 450
y 0 .-..---R----
phthalocyanine green mixtures in rutile Ti02
plus carbon black or burnt umber
0.25 % green + 0.1 % black
0.25 % green + 0.5 % burnt umber
0.25 % green+ 99.75 % Ti02
500 550 600
Wavelength (nm)
650 700
286
Figure A.24
Green mixtures (blue and yellow mixtures)
in rutile Ti0
2
.
Appendi x A
100
90
80
70
(ij 60
g
30
20
10
400
y 0
----- R----
green mixtures in rutile Ti0
2
5
.
.
.
. .
-- 1
/ "
/ . -
2 ;"" ./
---
-- 0.3 % phthalo blue+ 3.0% Permanent Yellow FGL
2 - - 0.2% phthalo blue+ 5.0% yellow iron oxide
3 - - 25.0% ultramarine blue+ 4.0% Permanent Yellow FGL
4 21.0% ultramarine blue+ 5.0% yellow iron oxide
5 0.5% phthalo green
450 500 550 600
Wavelength (nm)
650 700
Curves and Data for Pigments Used as Illustrative Problems 287
CIE Munsell
Figure Pigment composition x y
y
Hue Value/Chroma
A.25 0.025% CV, 99.975% R. Ti0
2
0.3020 0.3018 75.28 0.72 p 8.85/2.31
0.10% CV, 99.90% R. Ti0
2
0.2923 0.2844 61.40 9.54 PB 8.13/3.58
0.25% CV, 99.75% R. Ti0
2
0.2833 0.2680 49.63 0.33 p 7.44/5.34
0.625% CV, 99.375% R. Ti0
2
0.2717 0.2472 36.94 0.39 p 6.56/7.17
2.50% CV, 97.50% R. Ti0
2
0.2502 0.2086 19.44 9.93 PB 4.97/9.31
A.26 0.25% CV, 0.10% black, 99.65% R. Ti0
2
0.2854 0.2742 46.76 9.90 PB 7.25/4.64
0.25% CV, 0.50% B. umber, 99.25% R. Ti0
2
0.2963 0.2863 47.08 1.95 p 7.27/3.48
A.27 20.00% UB, 1.5% ROX, 78.5% R. Ti0
2
0.3015 0.2973 36.72 2.01 p 6.54/2.09
16.00% UB, 0.9% BON, 83.10% R. Ti0
2
0.2954 0.2786 38.66 3.14 p 6.69/4.16
0.30% PB, 1.25% QR, 98.45% R. Ti0
2
0.2780 0.2637 38.27 9.56 PB 6.66/5.50
Table A.6
CIE and Munsell notation (diffuse reflec-
y 0 -----R----
tance) for car b azo I e vi o I et (CV) and p u rp I e
1 00

mixtures (with ultramarine blue [UBJ,
red iron oxide [ROX], BON red, and
Quinacridone Red [QR]) .
Figure A.25
Carbazole dioxazine violet mixtures in
rutile Ti0
2
.
]j
g
Q)
u
c:
<ti
tJ
Q)
"" [!:'
+'
c:
Q)

Q)
Cl..
90
80
70
60
50
40
30
20
10
400 450
carbazole dioxazine violet mixtures in rutile Ti0
2
500 550
Wavelength (nm)
600 650 700
288
Figure A.26
Carbazole violet mixtures in rutile Ti0
2
,
and with carbon black or burnt umber
added.
Appendix A
90
80
70
(ij
...,
60
0
;::.
"'
u
c
(1j
...,
u
50
"'
;;:::

...,
c
"'
"'
40
CL
30
20
10
400
y
0
carbazole dioxazine violet mixtures in rutile Ti0
2
plus carbon black or burnt umber
, --.... '
I '

'
'
'
.. ..
. ..
. .
. .
: . '
. . '
.. '
. '
. '
'
--- 0.25% violet+ 0.1% black
0.25% violet+ 0.5% burnt umber
0.25% violet
450 500 550
Wavelength (nm)
600
R----
650 700
Figure A.27
Violet mixtures (red and blue mixtures) in
rutile Ti0
2
.
Curves and Data for Pigments Used as Illustrative Problems 289
y 0 14----R----
90
80
......... 4
. .
. ....
violet mixtures in rutile Ti0
2
. .
/ ... . ...
. .. .. 70
f1, ..... ..
I .. 3 I
: I '" .. I
:/ ' .... .:
: 2 ' ....
I
. .
: ,,.-- ' . :
60
fi I ' ' .:\ .... .. .. .....
50
'1 ' .. .. ..
.. '' .....
11 ,.
"I ~
~ - ,
40 I ~
30
20
10
400
1 --- 20.0% ultramarine blue + 1.5% red iron oxide
2 - - 16.0% ultramarine blue + 0.9% BON red
3 - - 0.3% phthalo blue + 1.249% Quinacridone Red
4 0.25% carbazole violet
450 500 550
Wavelength (nm)
600 650 700
290 Appendix A
Figure Pigment composition
A.28 1.00% MO, 99.00% R. Ti0
2
5.00% MO, 95.00% R. Ti0
2
25.00% MO, 75.00% R. Ti0
2
100.00% MO (masstone)
A.29 5.0% MO, 0.10% black, 94.9% R. Ti0
2
5.00% MO, 0.50% B. umber, 94.5% R. Ti0
2
A.30 1.50% BON, 2.5 % FGL, 96.00% R. Ti0
2
1.80% QR, 5.00% FGL, 93.20% R. Ti0
2
2.00% ROX, 2.00% YOX, 96.00% R. Ti0
2
Table A.7
CIE and Munsell notation (diffuse
reflectance) for molybdate orange (MO)
and orange mixtures (with BON red,
Permanent Yellow FGL, Quinacridone Red
[QR) , red iron oxide [ROX). and yellow
oxide [YOX]).
Figure A.28
Molybdate orange mi xtures in rutile Ti0
2
.
90
80
70
3
60
g
Q)
u
c

u
50 Q)
""'

.,._,
c
Q)

Q)
40
0..
30
20
10
400
CIE
x y
y
0.3399 0.3202 71 .98
0.3742 0.3229 54.99
0.4454 0.3292 34.03
0.6124 0.3471 15.25
0.3524 0.3206 48.92
0.3685 0.3257 51.17
0.3958 0.3375 47.44
0.4134 0.3556 47.02
0.3707 0.3354 49.33
molybdate orange mixtures in rutile Ti0
2
1.0%
5.0%
25.0 %
100.0%
450 500 550
Wavelength (nm)
Munsell
Hue Value/Chroma
4.16 R 8.68/ 3.32
4.40 R 7.7716.02
5.64 R 6.33/9.88
8.90 R 4.46/1 5. 75
4. 14 R 7.39/4.06
5.76 R 7.54/5.13
8.14 R 7.30/ 6.43
1.32 YR 7.27/6.86
9.1 5 R 7.42/4. 52
600 650 700
Figure A.29
Molybdate orange mixtures in rutile Ti0
2
,
and with carbon black or burnt umber
added.
Curves and Data for Pigments Used as Illustrative Problems
100
90
80
70
30
20
10
400
G
molybdate orange mixtures in rutile Ti0
2
plus carbon black or burnt umber
y
0
I
I
,
,
,
..
..
.
I .'
I .'
450
I .'
I .'
I .'
I.'
1:
1:
1:
5.0% orange + 0.1 % black
5.0% orange+ 0.5% burnt umber
5.0% orange
500 550
Wavelength (nm)
600
291
R----
""' ---- ----
;
............

650 700
292
Figure A.30
Orange mixtures (red and yellow mixtures)
in rutile Ti0
2
.
Appendix A
y 0 R----
90
80
70
60
g
<!)
u
c:

50
;:;:::

....,
c:
orange mixtures in rutile Ti0
2
3
40 /.----
... /
-----
30
2
,,,,,.-----
20
1 --- 1.5% BON red+ 2.5% Permanent Yellow FGL

2 - - 1.8% Quinacridone Red+ 5.0% Permanent Yellow FGL
3 - - 2.0% red iron oxide+ 2.0% yellow iron oxide
10
400
4 5.0% molybdate orange
450 500 550
Wavelength (nm)
600 650 700
Curves and Data for Pigments Used as Illustrative Problems 293
Figure Pigment composition
A.31 0.50% PB, 2.00% ROX, 97.50% R. Ti0
2
0.50% PG, 2.00% BON, 97.50% R. Ti0
2
2.50% YOX, 1.10% ROX, 21.2% UB,
75.2% R. Ti0
2
Table A.8
CIE and Munsell notation (diffuse
reflectance) for neutral mixtures (with
phthalo blue [PB], red iron oxide [ROX].
phthalo green [PG]. BON red , yellow oxide
[YOX]. and ultramarine blue [UBJ) .
90
80
70
"El 60
g
<IJ
u
c
.s
50
q::


c
<IJ

40
30
20
CIE
x y
y
Hue
Munsell
Value/Chroma
0.2837 0.2941
0.3089 0.2940
0.3107 0.3154
34.01
35.53
35.98
2.43 PB
6.83 p
2.27 RP
6.33/2.68
6.45/2.80
6.49/0.17
y 0
neutral mixtures in rutile Ti0
2
1 .......
- / 2""" I '
........... -.... . ' .. .
.............. I .. ./
/
.. ..:.:_ ................... : -.;-'
. -.. . ...... / '
: ............................ , ____ ":"<' /
: ............ ...... ___ ..,.,,,.
.: .................... ........
1 - - 0.5% phthalo blue+ 2.0% red oxide
2 -- 0.5% phthalo green+ 2.0% BON red
\
-..
4
3 2 .5 % yellow iron oxide + 1.1 % red iron oxide +
21.2% ultramarine blue
....................
10 4- 2.0% black
5 --- 0.5% black
Figure A.31 400 450 500
Neutral mixtures in rutile Ti0
2
.
550
Wavelength (nm)
600 650 700
%R
0.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
1.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
2.0
.1
.2
.3
.4
.5
.6
.7
.8
3
3.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
4.0
.1
2
.3
.4
.5
.6
.7
.8
.9
KIS
449.
249.
166.
124.
99.0
82.3
70.4
61.5
54.6
49.0
44.5
40.7
37.5
34.7
32.3
30.3
28.4
26.8
25.3
24.0
22.8
21.7
20.8
19.85
19.01
18.24
17.53
16.87
16.26
15.68
15.14
14.64
14.17
13.72
13.30
12.91
12.53
12.18
11.84
11.52
11.22
10.93
10.65
10.39
10.13
9.89
9.66
9.44
9.23
%R
5.0
. 1
.2
.3
.4
.5
.6
.7
.8
.9
6.0
.1
2
.3
.4
.5
.6
.7
.8
.9
7.0
.1
.2
.3
.4
.5
.6
.7
.8
3
8.0
.1
.2
.3
.4
.5
.6
.7
B
3
9.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
KJS
9.02
8.83
8.64
8.46
8.29
8.12
7.96
7.80
7.65
7.50
7.36
7.23
7.10
6.97
6.84
6.73
6.61
6.50
6.39
6.28
6.18
6.08
5.98
5.89
5.79
5.70
5.62
5.53
5.45
5.37
5.29
5.21
5.14
5.07
4.99
4 .93
4.86
4.79
4.73
4 .66
4.60
4.54
4 .48
4.42
4.37
4.31
4.26
4.20
4.15
4.10
%R
10.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
11.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
12.0
.1
2
.3
.4
.5
.6
.7
.8
.9
13.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
14.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
KIS
4 ~
400
3 ~
3.91
3.%
3 ~
3.77
3.n
3Y
3.M
3.60
3.56
3.52
3.48
3.44
3.41
3.37
3.33
3.30
3.26
3.23
3.19
3.16
3.13
3.09
3.06
3.03
3.00
2.97
2.94
2.91
2.88
2.85
2.83
2.80
2.77
2.74
2.72
2.69
2.67
2.64
2.62
2.59
2.57
2.54
2.52
2.50
2.48
2.45
2.43
%R
15.0
.1
.2
.3
.4
.5
.6
.7
8
.9
16.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
17.0
.1
2
.3
.4
.5
.6
.7
.8
.9
18.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
19.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
KIS
2.41
2.B
2.TI
2 ~
2.32
2.W
2.M
2 ~
2.M
2.22
2.21
2.19
2.17
2.15
2.13
2.11
2.10
2.08
2.06
2.04
2.03
2.01
1.993
1.977
1.961
1.945
1.929
1.913
1.898
1.883
1.868
1.853
1.838
1.824
1.809
1.795
1.781
1.767
1.754
1.740
1.727
1.713
1.700
1.687
1.674
1.662
1.649
1.637
1.624
1.612
%R
20.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
21.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
llD
.1
.2
.3
.4
.5
.6
.7
.8
.9
nn
.1
.2
.3
.4
.5
.6
.7
.8
3
24.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
KIS
1.600
1.588
1.576
1.565
1.553
1.542
1.530
1.519
1.508
1.497
1.486
1.475
1.465
1.454
1.443
1.433
1.423
1.413
1.403
1. 393
1.383
1.373
1.363
1.354
1.344
1.335
1.325
1.316
1.307
1.298
1.289
1.280
1.271
1.262
1.254
1.245
1.237
1.228
1.220
1.212
1.203
1.195
1.187
1.179
1.171
1.163
1.156
1.148
1.140
1.133
Appendix B
KIS versus %R Table
%R
25.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
26.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
27.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
28.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
29.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
K/ S
1.125
1.118
1.110
1.103
1.096
1.088
1.081
1.074
1.067
1.060
1.053
1.046
1.039
1.033
1.026
1.019
1.013
1.006
1.000
.993
.987
.981
.974
.968
.962
.956
.950
.944
.938
.932
.926
.920
.914
.908
.903
.897
.891
.886
.880
.875
.869
.864
.858
.853
.848
.842
.837
.832
.827
.822
%R
30.0
. 1
.2
.3
.4
.5
.6
.7
.8
.9
31.0
. 1
.2
.3
.4
.5
.6
.7
.8
.9
320
.1
.2
.3
.4
.5
.6
.7
.8
3
ED
.1
.2
.3
.4
.5
.6
.7
.8
3
34.0
. 1
.2
.3
.4
.5
.6
.7
.8
.9
KIS
.817
.812
.807
.802
.797
.792
.787
.782
.777
.773
.768
.763
.759
.754
.749
.745
.740
.736
.731
.727
.723
.718
.714
.710
.705
.701
.697
.693
.688
.684
.680
.676
.672
.668
.664
.660
.656
.652
.648
.644
.641
.637
.633
.629
.626
.622
.618
.614
.61 1
.607
%R
35.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
36.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
37.0
. 1
.2
.3
.4
.5
.6
.7
.8
.9
38.0
.1
2
.3
.4
.5
.6
.7
.8
.9
39.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
KIS
.604
.600
.596
.593
.589
.586
.583
.579
.576
.572
.569
.566
.562
.559
.556
.552
.549
.546
.543
.540
.536
.533
.530
.527
.524
.521
.518
.515
.512
.509
.506
.503
.500
.497
.494
.491
.488
.486
.483
.480
.477
.474
.472
.469
.466
.463
.461
.458
.455
.453
%R
40.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
41.0
.1
2
.3
.4
.5
.6
.7
.8
.9
42.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
43.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
44.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
KIS
.450
.447
.445
.442
.440
.437
.435
.432
.430
.427
.425
.422
.420
.417
.415
.412
.410
.408
.405
.403
.401
.398
.396
.394
.391
.389
.387
.385
.382
.380
.378
.376
.373
.371
.369
.367
.365
.363
.361
.358
.356
.354
.352
.350
.348
.346
.344
.342
.340
.338
%R
45.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
46.0
.1
.2
.3
.4
.5
6
.7
.8
.9
47.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
48.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
49.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
KIS
.336
.334
.332
.330
.328
.326
.325
.323
.321
.319
.317
.315
.313
.311
.310
.3M
3 ~
3 ~
.3m
3 ~
.299
.297
.295
.294
.292
.290
.288
.287
.285
.283
.282
.280
.278
.277
.275
.273
.272
.270
.269
.267
.265
.264
.262
.261
.259
.258
.256
.255
.253
.252
%R
50.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
51 .0
.1
2
.3
.4
.5
.6
.7
.8
.9

.1
.2
.3
.4
.5

.7
.8
9

.1
2
.3
.4
.5

J
B
9

.1
.2
.3
.4
.5
.6
.7
.8
9
KIS
.250
.249
.247
.246
.244
.243
.241
.240
.238
.237
.235
.234
.233
.231
.230
.228
.227
.226
.224
.223
.222
.220
.219
.218
.216
.215
.214
.212
.211
.210
.208
.207
.206
.205
.203
.202
.201
.1996
.1984
.1971
.1959
.1947
.1935
.1923
.191 1
.1899
.1888
.1876
.1864
.1853
%R
55.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
56.0
.1
.2
.3
.4
.5
.6
.7
.8
.9

.1
2
.3
.4
.5

J
B
9

.1
2
.3
.4
.5

.7
B
9
59.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
KIS
.1841
.1829
.1818
.1807
.1795
.1784
.1773
.1762
.1751
.1740
.1729
.1718
.1707
.1696
.1685
.1675
.1664
.1653
.1643
.1632
.1622
.1612
.1601
.1591
.1581
.1571
.1561
.1551
.1541
.1531
.1521
.1511
.1501
.1491
.1482
.1472
.1462
.1453
.1443
.1434
.1425
.1415
.1406
.1397
.1388
.1378
.1369
.1360
.1351
.1342
%R
60.0
.1
.2
.3
.4
.5
.6
.7
.8
.9

.1
2
.3
.4
.5

J
B
9

.1
2
.3
.4
.5

J
B
9

.1
2
.3
.4
.5

.7
.8
9
64.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
KIS
.1333
.1325
.1316
.1307
.1298
.1290
.1281
.1272
.1264
.1255
.1247
.1238
.1230
.1222
.1213
.1205
.1197
.1189
.1181
.1173
.1165
.1157
.1149
.1141
.1133
.1125
.1117
.1110
.1102
.1094
.1087
.1079
.1071
.1064
.1056
.1049
.1042
.1034
.1027
.1020
.1013
.1005
.0998
.0991
.0984
.0977
.0970
.0963
.0956
.0949
%R
65.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
66.0
.1
.2
.3
.4
.5
.6
.7
.8
.9

.1
2
.3
.4
.5

J
B
9

.1
.2
.3
.4
.5
.6
.7
B
9

.1
2
.3
.4
.5

J
B
9
KIS
.0942
.0936
.0929
.0922
.0915
.0909
.0902
.0895
.0889
.0882
.0876
.0869
.0863
.0857
.0850
.0844
.0838
.0831
.0825
.0819
.0813
.0807
.0801
.0794
.0788
.0782
.0776
.0771
.0765
.0759
.0753
.0747
.0741
.0736
.0730
.0724
.0719
.0713
.0707
.0702
.0696
.0691
.0685
.0680
.0675
.0669
.0664
.0659
.0653
.0648
%R
70.0
.1
2
.3
.4
.5
.6
.7
.8
.9
71.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
72.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
73 .0
.1
.2
.3
.4
.5
.6
.7
.8
.9
74.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
K/ S
.0643
.0638
.0633
.0627
.0622
.0617
.0612
.0607
.0602
.0597
.0592
.0587
.0583
.0578
.0573
.0568
.0563
.0559
.0554
.0549
.0544
.0540
.0535
.0531
.0526
.0522
.0517
.0513
.0508
.0504
.0499
.0495
.0491
.0486
.0482
.0478
.0474
.0469
.0465
.0461
.0457
.0453
.0449
.0445
.0440
.0436
.0432
.0428
.0425
.0421
%R
75.0
.1
2
.3
.4
.5
.6
.7
.8
.9
76.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
77.0
.1
.2
.3
.4
.5
.6
.7
.8
.9

.1
2
.3
.4
.5

J
B
9

.1
2
.3
.4
.5

J
B
9
KIS
.0417
.0413
.0409
.0405
.0401
.0398
.0394
.0390
.0386
.0383
.0379
.0375
.0372
.0368
.0365
.0361
.0357
.0354
.0350
.0347
.0344
.0340
.0337
.0333
.0330
.0327
.0323
.0320
.0317
.0314
.0310
.0307
.0304
.0301
.0298
.0294
.0291
.0288
.0285
.0282


mn

mg
fil64




%R
80.0
.1
.2
.3
.4
.5
.6
.7
.8
.9

.1
.2
.3
.4
.5

.7
.8
9

.1
2
.3
.4
.5
.6
.7
.8
9

.1
2
.3
.4
.5

J
B
9
84.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
KIS
.0250
.0247
.0244
.0242
.0239
.0236
.0234
.0231
.0228
.0226
.0223
.0220
.0218
.0215
.0213
.0210
.0208
.0205
.0203
.0200
.0198
.0195
.0193
.0190
.0188
.0186
.0183
.0181
.0179
.0176
.0174
.0172
.0170
.0167
.0165
.0163
.0161
.0159
.0157
.0155


mq
m%
m44


.0138
.0136

%R
85.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
86.0
.1
.2
.3
.4
.5
.6
.7
.8
.9

.1
.2
.3
.4
.5
.6
.7
.8
9
RO
.1
2
.3
.4
.5

J
B
9
89.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
KIS
.0132
.0130
.0129
.0127
.0125
.0123
.0121
.0119
.0118
.0116
.0114
.0112
.0111
.0109
.0107
.0105
.0104
.0102
.0100
.00987
.00971
.00955
.00939
.00924
.00908
.00893
.00878
.00863
.00848
.00833
.00818
.00804
.00789
.00775
.00761
.00747
.00733
.00720
.00706
.00693
.00680
.00667
.00654
.00641
.00628
.00616
.00604
.00591
.00579
.00567
%R
90.0
.1
2
.3
.4
.5
.6
.7
.8
.9
91.0
.1
.2
.3
.4
.5
.6
.7
.8
.9

.1
2
.3
.4
.5

J
B
9

.1
2
.3
.4
.5

J
B
9
94.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
KIS
.00556
.00544
.00532
.00521
.00510
.00499
.00488
.00477
.00466
.00456
.00445
.00435
.00425
.00415
.00405
.00395
.00385
.00376
.00366
.00357
.00348
.00339
.00330
.00321
.00313
.00304
.00296
.00287
.00279
.00271
.00263
.00256
.00248
.00241
.00233
.00226
.00219
.00212
.00205
.00198
.00192
.00185
.00179
.00172
.00166
.00160
.00154
.00148
.00143
.00137
%R
95.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
96.0
. 1
.2
.3
.4
.5
.6
.7
.8
.9
97.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
98.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
99.0
.1
.2
.3
.4
.5
.6
.7
.8
.9
100.0
KIS
.00132
.00126
.00121
.00116
.00111
.00106
.00101
.00097
.00092
.00088
.00083
.00079
.00075
.00071
.00067
.00064
.00060
.00056
.00053
.00050
.00046
.00043
.00040
.00038
.00035
.00032
.00030
.00027
.00025
.00023
.00020
.00018
.00017
.00015
.00013
.00011
.00010
.00009
.00007
.00006
.00005
.00004
.00003
.00002
.00002
.00001
.00001
.00000
.00000
.00000
0.00000

<

c:
"'
oft
:>o
;;;4

i>"
"" \0
1.11
Appendix C
Colour Index Name
The standard method for naming pigments and dyes worldwide is by
means of their Colour Index generic name and number, plus the five-digit
constitution number of each pigment or dye. This system for characteriz-
ing pigments and dyes is the result of a continuing joint effort of the
Society of Dyers and Colourists of Great Britain and the American
Association of Textile Chemists and Colorists in the United States. The
Colour Index generic name is based on the most common method by
which the colorant is applied, such as vat dyes (Vat), acid dyes (Acid),
and so on. All pigments are in a single generic class-pigments (P) - with
the exception of a few based on vat dyes that may be used in place of the
pigment classification. There are twenty generic classes for dyes, assigned
depending on their use. The hue description follows the use description.
Among pigments there are nine hue descriptions: blue (B), green (G), vio-
let (V), red (R), orange (0 ), brown (Br), yellow (Y), white (W), and
black (Bk). The generic number following the name refers to the particu-
lar type of pigment. Thus, the Colour Index generic name and number of
natural BaS0
4
, barytes, is Pigment White 22 (PW 22), whereas the syn-
thetic BaS0
4
, blanc fixe, has the generic name and number Pigment
White 21 (PW 21 ).
A complete Colour Index designation will also contain a five-
digit constitution number after the generic name and number to indicate
the chemical constitution of each colorant when it is known. (Some mod-
ern pigments have no constitution number because their exact chemical
structures are proprietary. ) Blocks of numbers have been assigned accord-
ing to the general chemical character of the colorant. Thus, all natural
pigments are given Colour Index constitution numbers from 7 5000 to
75999; inorganic pigments are assigned numbers from 77000 to 77999;
and phthalocyanine colorants are numbered 74000 to 74999. In the case
of the barium sulfates, both the barytes and blanc fixe have a constitu-
tion number of 77120, because both consist of the compound BaS04,
even though one is of natural origin.
In appendix D, which describes a method for indexing spec-
trophotometric reflectance curves and gives a representative list of red
pigments as an example, the pigments are designated by the generic name
Colour Index Name 297
PR (Pigment Red), followed by their generic number. Note that some
have a further notation to describe the metallizing ion used: for example,
PR 63 Mn and PR 63 Ca describe the particular BON red version (man-
ganese or calcium, respectivel y) .
For a more complete discussion of the Colour Index,
Billmeyer and Saltzman (1981 ) may be consulted.
Appendix D
Compilation of Spectral Reflectance Curves of Pigments:
Representative Example
The need for a reference library of pigment reflectance curves in the
visible region has been emphasized many times throughout this mono-
graph; its value to the spectrophotometric techniques described in the
preceding chapters should be readily apparent. This appendix describes
how such a coll ection could be organized and presents an example in
terms of some of the many red pigme nts available today. The method
illustrated here of organizing curves on the basis of the number and the
wavelength of the absorption bands is an adaptation of the method for
providing a curve-shape index for identification outlined by Shurcliff
(1942). Ideall y, curves plotted as log 1/(K/S) versus wavelength, instead
of reflectance versus wavelength, would be more useful, as described in
the section on colorant calculations and identifications in chapter 4.
This derived absorption curve shape is independent of the concentra-
tion. However, obtaining such curves would require computation of the
values from the reflectances. Hence, collecting the reflectance curves is
needed as the first task.
If a complete library were to be done, the first factor form-
ing the basis of the arrangement would be hue: The pigments would
be arranged according to their absorption regions, from long to short
wavelength-that is, greens (which absorb in the red region ), then blues,
purples, reds, oranges, and finally yellows (which absorb in the short-
wavelength violet region). There would be a separate category for
neutrals, whites, and blacks. Each pigment would be identified by its
Colour Index name (see appendix C) and by its common chemical -type
description. If a curve has been identified by a manufacturer's trade
name or product number (or both), that too would be included because
of its specificity.
Where available, masstones would be included, but in gen-
eral, mixtures of all the absorbing pigments with whi te provide the most
information about curve shapes. Their characteristic absorption bands
apply to samples at incomplete hiding as well, such as might be seen in
watercolors, printing inks, and textile printing.
Familiar historic pigment names may not be included unless
their composit ion is firml y established. The reader is warned again that
pigment names used on tubes of artists' colors are not necessarily those
of the actual pigments. For example, Johnston and Feller (1963) analyzed
Compilation of Spectral Reflectance Curves of Pigments: Representative Example 299
four tubes of artists' watercolors labeled "indigo." Only one tube
(American) contained the real indigo pigment. The others contained a
diversity of pigments such as phthalocyanine blue and black; iron blue
plus ultramarine blue and black; and ultramarine blue, phthalocyanine
green, phthalocyanine blue, pyrazolone red, and black! Thus, even when
artists record the palettes they use, one must beware. The danger of mis-
labeling, illustrated by this example for indigo, is a valid possibility dat-
ing from the times when artists purchased colorants predispersed by the
supplier of artists' materials.
Even in previous centuries, when artists often dispersed the
pigments themselves, pigment names still cannot be considered to be
representative of pigment composition. Feller and Johnston-Feller
( 1997), for example, found discrepancies in the meaning of the color
name Vandyke Brown. To the English and Germans, the name referred
to the humic-earth pigment that is mined in peat bogs near Cologne,
Germany, and other places, and which is a natural, organic pigment,
relatively transparent, and of marginal lightfastness. To the French,
however, the name Vandyke Brown referred to an iron oxide pigment
that is more opaque and of high lightfastness. For the humic-earth pig-
ment, they used the name Cologne Earth, or some other name derived
from the pigment's origin.
In the organizational scheme for a library of reflectance
curves illustrated here, the second category is the number of absorption
bands present in the curves. The most simple curves-the yellows, for
example-may exhibit no absorption maximum in the visible region.
Other pigments may exhibit only a single absorption maximum. Even
if such cases are encountered, the curve shape can tell the pigment-
identification sleuth which pigments may and may not be present.
The number of absorption maxima can range from none (or
simply a shoulder-that is, a slight dip in the curve) to one absorption
band, then two absorption bands, and so on. When two or more absorp-
tion bands are present, the curves are categorized on the basis of the
wavelength of the major absorption band.
Table D.l is an example of this organizational scheme
applied to a representative list of red pigments whose curves are found
in figures D.l-D.26. The major headings across the table are as follows:
1. Colour Index name
2. Type (the commonly used chemical-type description)
3. Wavelength of absorption bands (entries in the table list
the wavelength of the major band first, followed by the
wavelengths of any progressively minor absorption bands,
and end with the KIS ratio relative to the major band; the
ratio is in parentheses)
4. Figure (number of the figure illustrating the reflectance
curve)
Entries in the table are grouped according to their hue and
number of absorption bands-for example, "RED: Three Absorption
300
Colour
Index name
PR 48:1
PR 38
PR 101
PR 104
PR 4
PR 63 Mn
PR 49 Ca
PR 63 Ca
PR 23
PR 49 Ca
PR 49 Ba
PR 48:2
PR 90
PR 49 Na
PR 112
PR 22
PR 48:3
PR 166
PR 38
PR 48:1
PR 63
PR 41
PR 87
PR 1
PR 179
PR 177
PR 190
Vat R 29
PR 197
PR 168
PR 207
PR 123
Appendix D
Wavelength
Type of absorption bands
RED: One Absorption Band
red 2B (barium)
pyrazolone in acrylic
synthetic red oxide
molybdate orange
orthoch I orparan itran i Ii n e
555
540
545
530
520
RED: Two Absorption Bands
BON maroon (manganese) 585, 540 (0.89)
lithol maroon (resinated) 580, 545 (0.89)
(calcium)
BON red (calcium) 560, 525 (0 89)
naphthol red 570, 535 (0.82)
lithol red (resinated) 560, 530 (0.88)
(calcium)
lithol red (resinated) 560, 530 (0.88)
(barium)
red 2B (calcium) 560, 515 (0 65)
phloxine 555, 510 (0.78)
lithol red (resinated) 545, 510 (0.84)
(sodium)
naphthol 550, 512 (0.92)
naphthol 550, 515 (0 88)
red 2B (strontium) 550, 515 (0.72)
disazo condensation 545, 512 (0.90)
pyrazolone in nitrocellulose 540, 575 (0 84)
red 2B (barium) 540, 505 (0.72)
BON red (Ba-Mn) 530, 570 (0.94)
dianisidine 528, 565 (0.90)
thioindigo 525, 550 (0.61 )
paranitraniline 535, 575 (0 76)
perylene maroon 475, 520 (0.82)
RED: Three Absorption Bands
anthraquinoid
perylene
I ndofast Brilliant
Scarlet
brominated pyranthrone
orange
brominated anthanthrone
orange
Quinacridone Red
565, 535 (0.92), 455
(0.48)
565, 527 (0.85), 480
(0.92)
563, 528 (0.85), 483
(0.97)
562, 477 (0.61), 450
(0.59)
545, 520 (0.94), 485
(0.68)
509, 535 (0.91 ), 475
(0.50)
RED: Four Absorption Bands
perylene vermilion 555, 510 (0.81 ), 480
(0.66), 445 (0.39)
' The figures in parentheses are the K/S ratios of the lower absorptions to the major absorption.
Table D.1
Representative list of red pigments.
Figure
D.2
D.1
D.4
D.3
D.5
D.7
D.6
D.7
D.8
D.6
D.6
D.9
D.12
D.6
D.1 1
D.1 3
D.10
D.15
D.14
D.16
D.7
D.17
D.19
D. 18
D.21
D.20
D.21
D. 22
D.23
D.24
D.25
D.26
D.21
Compilation of Spectral Reflectance Curves of Pigments: Representative Example 301
Bands." The major absorption bands are ordered from long to short
wavelength, in keeping with the qualitative identification procedure rec-
ommended in chapter 4. This may seem unnatural in view of the left-to-
right reading habit generally applied to the interpretation of graphs.
However, for the purpose described, and with a little practice, this pro-
cedure works best.
All of the curves used for table D. l are based on manufactur-
ers' literature. They are, therefore, modern pigments manufactured after
1960. A representative entry under the grouping "RED: Three Absorption
Bands" is PR 190; perylene; 565, 527 (0.85), 480 (0. 92); D. 21.
One decision that may be subject to question in this organiza-
tional scheme is the choice of the hue category. When does a pigment
become a red rather than an orange? Or when does a reddish pigment
become a violet? Since the categorization of colors close in hue is arbi-
trary at times, the curves are placed in hue categories according to their
assigned Colour Index names. Thus, some "oranges" are classified as
reds, and some "reds" are listed with the violets, and so on, in accor-
dance with their designation in the Colour Index.
The reader should be aware that the numbers for the wave-
lengths and for the reflectances, from which the KIS ratios were calcu-
lated, were obtained by reading them off the curves shown in this
appendix and are, therefore, not precise. The wavelengths were
determined only to the nearest 10 nm; the reflectances only to the
nearest full percent.
Another part of this organizational scheme that is subjec-
tive-and thus open to question-is the choice of the wavelength that
should be used to identify broad shoulders and absorption regions. An
example is curve D.14 for PR 38. The reader may not agree with the
wavelength assignments listed in table D.l. If an atlas of curves is ever
prepared, the editors will be confronted with such decisions.
Accumulating a library of spectral reflectance curves must be
regarded as an ongoing task, to be added to as additional references or
new materials are collected. The accompanying list of red pigments
serves only as an introductory demonstration of the possibilities.
In the appendix D figures, the colorant amount in the
colorant-Ti0
2
ratio is indicated by the letter C- for example, 25 C/75
Ti0
2
. The curves illustrated in figures D.1, D.5-D.8, D.12-D.14, D.16,
D.21, D.23, and D.24 are redrawn from illustrations in volume 1 of the
first edition of the Pigment Handbook (Patton 1973 ). The others are
redrawn from the manufacturers' literature accumulated by the author
and from curves prepared by the author. Sources represented are Ciba
Specialty Chemicals (formerly Ciba-Geigy), DuPont, and Harmon Colors.
302
Figure D.1 (from Patton 1973. Copyright
1973 John Wiley & Sons, Inc. Reprinted
by permission).
Appendix D
100
90
80
70
<ii
+"'
60
g
Q)
u
c
.;g
u
50
Q)
;:;:::
~
+"'
c
Q)
~
Q)
40
CL
30
20
10
400 450
y 0
Caddy Red Toner
pyrazolone red in acrylic lacquer
(PR 38)
500 550 600
Wavelength (nm)
~ R
650 700
Figure D.2
Compilation of Spectral Reflectance Curves of Pigments: Representative Example
90
80
70
.,
60
"' :J

'.,
"'
u
c
1l
50
u
"' '+=

.._,
c
"'
u
Q:; 40
a_
30
20
10
400 450
y
0
Permanent Red 28 (barium) (PR 48:1)
500 550
Wavelength (nm)
600
303
R----
650 700
304 Appendix D
100
90
80
70
<il 60
0
;::.
<lJ
u
c
<1l
....
u
50
<lJ
q::

....
c
<lJ
u
Q:;
40 CL
30
20
10
400
Figure D.3
G
y 0
X-2806 Rex Orange
molybdate orange (PR 104)
450 500 550
Wavelength (nm)
600 650 700
Figure D.4
Compilation of Spectral Reflectance Curves of Pigments: Representative Example 305
100
90
80
70

60
:J


Q)
u
c

50
u
Q)
;:;::::

.....
c
Q)
u
Q:; 40
CL
30
20
10
400 450 500
y 0
synthetic red iron oxide
(PR 101)
550
Wavelength (nm)
600
-----R----
650 700
306
Figure D.5 (from Patton 1973. Copyright
1973 John Wiley & Sons, Inc. Reprinted
by permission).
Appendi x D
90
80
70
<il 60
....,
g
QJ
u
c
!l
u
50
QJ
q::

....,
c
QJ
u
v
40
CL
30
20
10
400
Chlorinated Para Red (PR 4)
orthochlorparanitraniline red
50 C/50 Ti02
450 500 550 600
Wavelength (nm)
650 700
Figure 0.6 (from Patton 1973. Copyright
1973 John Wiley & Sons, Inc. Reprinted
by permission).
Compilation of Spectral Reflectance Curves of Pigments: Representative Example
100
90
80
70

60
:J


<lJ
u
c:
_:g
50
u
<lJ
;:;:::


c:
<lJ

40 <lJ
0..
30
20
10
400
Lithol Red (PR 49)
450 500
G y 0
Ca lithol
Ca maroon
550
Wavelength (nm)
600
307
650 700
308
Figure D.7 (from Patton 1973. Copyright
1973 John Wi ley & Sons, Inc. Reprinted
by permission).
Appendix D
90
80
70
<il 60
g
"'
u
c
.is
u
50
"'
q::

....,
c
"'
"'
40
CL
30
20
10
400 450 500
y 0
BON maroon (PR 63)
1 C/25 Ti02 -._.
masstone
/
550
Wavelength (nm)
600
R----
650 700
Figure 0.8 (from Patton 1973. Copyright
1973 John Wiley & Sons, Inc. Reprinted
by permission).
Compilation of Spectral Reflectance Curves of Pigments: Representative Example
-V---l ... --
90
80
70
Q)
60
"' :J
~
:s
"'
u
c
cu
50
.....
u
"'
<;:::
!'::'
.....
c
"'
40
"'
Cl..
30
20
10
400 450
B _ . ~
G y 0
naphthol type (PR 23)
33 C/67 Ti0
2
500 550
Wavelength (nm)
600
309
R----
650 700
310 Appendix D
90
80
70
-;;i 60
g
"'
u
c:
.:'l
u
50
"'


.....
c:
"'
"'
40
CL
30
20
10
400
Figure D.9
450 500
y 0
X-2277 Monterey Red,
red 28-calcium (PR 48:2)
50 C/50 Ti02
550
Wavelength (nm)
600
14----- R ----
650 700
Figure D.10
Compilation of Spectral Reflectance Curves of Pigments: Representative Example
-v
100
90
80
70
60
g
Q)
u
c
.l'l
u
50
Q)
q::

......
c
Q)

Q)
40 a_
30
20
10
400
B
X-3566 Adirondack Red,
red 28-strontium
(PR 48:3)
450 500
G
y
50 C/50 Ti02
550
Wavelength (nm)
0
600
311
R
650 700
312 Appendix D
100
90
80
70
<ii
.....
60
g
"'
u
c
.r:l
u
50
"'
;;:::

.....
c
"'
u
Q:;
40 CL
30
20
10
400
Figure D.11
450 500
y 0
X-3079 Waco Red
(PR 112) (naphthol)
50 C/ 50 Ti0
2
550
Wavelength (nm)
600 650 700
Figure D.12 (from Patton 1973. Copyright
1973 John Wiley & Sons, Inc. Reprinted
by permission).
Compilation of Spectral Reflectance Curves of Pigments: Representative Example 313
90
80
70

60
:::i


Q)
u
c
rd
50
t>
Q)
;:;:::

......
c
Q)
u
OJ 40
CL
30
20
10
400 450 500
y 0
Phloxine Pigment
(PR 90)
550
Wavelength (nm)
600
-----R----
650 700
314
Figure D.13 (from Patton 1973. Copyright
1973 John Wiley & Sons. Inc. Reprinted
by permission).
Appendix D
90
80
70
Q)
60
V\
:::l
~
~
<!)
u
c
_:g
50
u
<!)
;+::
~
.....
c
<!)
~
40 <!)
Q_
30
20
10
400
y 0
X-3079 naphthol type (PR 22)
33 C/ 67 Ti02
= ~
450 500 550
Wavelength (nm)
600
R----
650 700
Figure D.14 (from Patton 1973. Copyright
1973 John Wiley & Sons, Inc. Reprinted
by permission).
Compilation of Spectral Reflectance Curves of Pigments: Representative Example
90
80
70
"B
60
g
cu
u
c
<d
u
50
cu
;:;:::

_.._,
c
cu

cu
40
0..
30
20
10
400 450
y
0
Pyrazolone Red C Toner in
nitrocellulose lacquer (PR 38)
500
50 C/50 Ti0
2
550
Wavelength (nm)
600
315
R----
650 700
316 Appendix D
90
80
70

60
g
Q)
u
c
<tj
t)
50
Q)
<+:

c
Q)

Q)
40
"-
30
20
10
400
Figure D.15
450
y 0
Cromophtal Scarlet Red
disazo condensation (PR 166)
500 550
Wavelength (nm)
600
i.---- R ----
650 700
Figure 0.16 (from Patton 1973. Copyright
1973 John Wiley & Sons, Inc. Reprinted
by permission).
Compilation of Spectral Reflectance Curves of Pigments: Representative Example
90
80
70
]
60
g
"'
u
c:

u
50
"'
<;::

.._,
c:
"'
"'
40 CL
30
20
10
400 450
X2478 Orono Red, red
2B-barium (PR 48:1)
500 550
Wavelength (nm)
600
317
650 700
318 Appendix D
100
90
80
70
!ii 60
g
Q)
u
c
(1j
....,
u
50 Q)
;:;:::

....,
c
Q)
u
v
40 Q._
30
20
10
400
Figure D.17
450 500
y 0
Vulcan Fast Red
(PR 41 )(dianisidine red)
50 C/50 Ti0
2
550
Wavelength (nm)
600
-----R----
masstone
650 700
Figure D.18
Compilation of Spectral Reflectance Curves of Pigments: Representative Example 319
90
80
70
<ii 60
0
;::.
"'
u
c

u
50
"'
""'
1:'
....,
c
"'
"'
40 CL
30
20
10
400 450 500
y
Para Red, Dark
paranitraniline red (PR 1)
550
Wavelength (nm)
0
600
-----R----
masstone
650 700
320 Appendix D
100
90
80
70
OJ
60
"' ::J

:s
"'
u
c
<rj
50
tJ
"'
""

....,
c
"'
u
:;; 40
CL
30
20
10
400
Figure D.19
450 500
y 0
thioindigo red (PR 87)
550
Wavelength (nm)
600 650 700
Figure D.20
Compilation of Spectral Reflectance Curves of Pigments: Representative Example 321
100
90
80
70
l'i
60
g
"'
u
c:
fl
u
50
"'
;:;:::

....,
c:
"'
u
Q:;
40
CL
30
20
10
400 450 500
G
y
Chromophtal Red 3B
anthraquinoid (PR 177)
550
Wavelength (nm)
0 14---R----
600 650 700
322
Figure D.21 (from Patton 1973. Copyright
1973 John Wiley & Sons, Inc. Reprinted
by permission).
Appendix D
100
90
80
70
di
60
"' ::J

:s
Q)
u
c

50
u
Q)
;;:::
e:
...,,
c
Q)

40 Q)
CL
30
20
10
400
y 0
(Perylene Vermilion)
/
PR 123
(Perylene Maroon)
450 500 550 600 650 700
Wavelength (nm)
Figure D.22
Compilation of Spectral Reflectance Curves of Pigments: Representative Example
v
Vl
:::J


cu
u
c
<U
.....
u
cu
;;:::
1:'
.....
c
"'
"'
Cl..
-V --1-IE--- B ---.14--- G y 0
100
90
80
70
60
50
40
30
20
10
400 450
lndofast Brilliant Scarlet
(derived from vat red 29)
500 550
Wavelength (nm)
600
323
650 700
324
Figure D.23 (from Patton 1973. Copyright
1973 John Wiley & Sons, Inc. Reprinted
by permission).
Appendix D
100
90
80
70
Q)
60
V>
:::J


Q.)
u
c:

50
u
Q.)
""' I:'
.....
c:
Q.)

40 Q.)
(l_
30
20
10
400
G y 0
Brominated Pyranthrone Orange
(PR 197)
450 500 550
Wavelength (nm)
600
-----R----
650 700
Figure D.24 (from Patton 1973. Copyright
1973 John Wiley & Sons, Inc. Reprinted
by permission).
Compilation of Spectral Reflectance Curves of Pigments: Representative Example 325
100
90
80
70
Q)
60
V\
::>
:t:

"'
u
c
.s
50
u
"'
;;:::

.._,
c
"'
40
"'
a_
30
20
10
400 450
G
y 0
Brominated Anthanthrone Orange
(PR 168)
500 550
Wavelength (nm)
600
-----R----
650 700
326 Appendix D
90
80
70

60
:J

:s
cu
u
c
_:g
50
u
cu
<;:::

c
cu

40 cu
CL
30
20
10
400
Figure 0 .25
450 500
y 0
Quinacridone Red
(PR 207)
550
Wavelength (nm)
600
-----R--- -
650 700
Figure D.26
Compilation of Spectral Reflectance Curves of Pi gments: Representative Example
90
80
70

60
::::>

2
<lJ
u
c
_::g
50
u
<lJ


_...,
c
<lJ

40 <lJ
CL
30
20
10
400 450
lndofast Brilliant Scarlet Toner
Perylene Vermilion (PR 123)
500 550
Wavelength (nm)
600
327
650 700
Appendix E
Recommended Reading
Many of the books and articles listed below have been referred to in the
text; a few have not. They are categorized here by subject; notes explain
their major emphasis.
General Textbooks on Color Science
Berger-Schunn, A.
1994 Practical Color Measurement. Original in German, translated by the
author and Max Saltzman. New York: John Wiley and Sons.
Brief coverage of the essentials of color perception, color measurement,
correlations between reflectance and colorant concentration, colorant
st rength, and computer color matching. The stated objective is to describe
"only what every user ... needs to know." The viewpoint is oriented
toward textiles.
Billmeyer, F. W., Jr., and M. Saltzman
1981 Principles of Color Technology, 2d ed. (1st ed., 1966). New York: John
Wiley and Sons.
Judd, D. B.
The best elementary book available. The third edition of this work
appeared in 2000 (New York: John Wiley and Sons ), revised and updated
by Roy S. Berns.
1952 Color in Business, Science, and Industry, lst ed. New York: John Wiley
and Sons.
The most comprehensive textbook available. Comparison of this first edi-
ti on of Judd's with subsequent editions prepared with Wyszecki provides
an interesting histor y of developments in color science and application
over nearly twenty-five yea rs of extensive research and application, as
well as an appreciation of the constancy of basic ideas.
Judd, D. B., and G. Wyszecki
1975 Color in Business, Science, and Industry, 3d ed. (2d ed., 1963). New
York: J oh n Wiley and Sons.
Wright, W. D.
1969 The Measurement of Colour, 4th ed. New York: Van Nostrand.
A classic.
Recommended Reading 329
Wyszecki, G. , and W. S. Stiles
1982 Color Science: Concepts and Methods, Quantitative Data and Formulae,
2d ed. (1st ed., 1967). New York: John Wiley and Sons.
Both editions are superb sources of the methods, data, and reference
materials that make up the science of colorimetry. Encyclopedic in scope.
Spectrophotometry and Its Applications
Frei, R. W., and J. D. MacNeil
1973 Diffuse Reflectance Spectroscopy in Environmental Problem Solving.
Cleveland: CRC Press, Division of the Chemical Rubber Co.
An interesting book that describes different applications of spectropho-
tometry and the Kubeika-Munk equation.
Hunt, R. W. G.
1982 Measuring Colour, 1st ed. (2d ed., 1991 ). Chichester, Eng.: Ellis Horwood.
Kortiim, G.
Provides the basic facts needed to measure color. It is lucidl y written and
eminently suitable as an up-to-date textbook.
1969 Reflectance Spectroscopy: Principles, Methods, Applications. Translated
from the German by James E. Lohr. New York: Springer-Verlag.
Discusses specular and diffuse reflection, and absorption and light scatter-
ing according to Rayleigh, Mie, and Kubeika and Munk.
Stearns, E. I.
1969 The Practice of Absorption Spectrophotometry. New York: John Wiley
and Sons.
Volz, H. G.
1995
Presents many of the methods and techniques and practical applications of
absorption spectrophotometry, particularly of solutions. It is highly worth-
while for its emphasis on the understanding of and logical approach to the
analysis of multicomponent materials based on basic principles not requir-
ing elegant computer facilities. Also steps the reader through the prepara-
tion of a computer program for color matching. Dyestuffs orientation.
Industrial Color Testing: Theory and Methods. Engli sh translation.
Deerfield Beach, Fla.: VCH.
This is an excellent book on the relation of color to spectra; the theory of
light scattering and absorption by pigments; and discussions of the
Beer-Bouguer law, the Kubeika-Munk equation, Mie theory, and related
measurement and test methods.
Appearance
Hunter, R. S., and R. W. Harold
1987 The Measurement of Color and Appearance, 2d ed. New York: John
Wi ley and Sons.
330 Appendi x E
Colorant Formulation Calculations
Allen, E.
1980 Colorant formulations and shading. In Optical Radiation Measurements,
ed. F. Grum and C. J. Bartleson, 2:289-336. New York: Academic Press.
An excellent summary and review.
Davidson, H. R., and H. Hemmendinger
1966 Color prediction using the two-constant turbid-media theory. Journal of
the Optical Society of America 56:1102-9.
Kuehni, R. G.
1975 Computer Colorant Formulation. Lexington, Mass.: D. C. Heath and Co.
Oriented toward textiles but includes details for computing
Kubeika-Munk two-constant-theory colorant formulations.
Methods and Techniques
American Societ y for Testing and Materials
ASTM Standards on Color and Appearance Measurements. West
Conshohocken, Pa.: American Society for Testing and Materials.
Revised periodically. Contains eighty or more methods for measuring
color and appearance attributes, as well as a list of simi lar length of
auxiliary ASTM methods not included but published elsewhere.
Celikiz, G., and R. G. Kuehni, eds.
1983 Color Technology in the Textile Industry. Research Triangle Park, N.C.:
American Association of Textile Chemists and Colorists.
A collection of articles of general interest, some original and some previ-
ous ly published. Prepared under the aegis of AATCC Committee RA36,
Color Measurement Test Methods.
Sensation and Perception
Boynton, R. M.
1979 Human Color Vision. New York: Holt, Rinehart and Winston.
An excellent, comprehensive, but very readable text on human vision.
The author states in the preface that he hopes it will be of value to non-
specialists, such as "artists, interior decorators, illuminating engineers,
architects, graphic arts designers," and many others, as well as to stu-
dents studying sensation and perception.
Evans, R. M.
1948 An Introduction to Color. New York: John Wiley and Sons.
Out of print but worth looking for. Beautifully illustrated camera
techniques.
Recommended Reading 331
1971 The perception of color. In Industrial Color Technology, ed. R. F. Gould,
Advances in Chemistry Series 107:43-68. Washington, D.C.: American
Chemical Society.
Paper presented at a symposium, arranged by Ruth M. Johnston and Max
Saltzman, during the 156th Annual Meeting of the American Chemical
Society. It contains beautiful color illustrations, never published else-
where, of perceptual color effects.
1974 The Perception of Color. New York: John Wiley and Sons.
General
Minnaert, M. G. J.
1993 Light and Color in the Outdoors. Translated and revised by Len Seymour.
New York: Springer-Verlag.
This revision of Minnaert's original book contains 223 figures and 49
color plates in a hard copy that is a delight to read, look at, and enjoy.
O'Neill, M., and L. Weisgard
1961 Hailstones and Halibut Bones. Garden City, N.Y.: Doubleday and Co.
A charming book of poems about each color for children of all ages,
beautifully illustrated in color by Weisgard.
Optical Society of America, Committee on Colorimetry
1953 The Science of Color. New York: Thomas Y. Crowell Co.
A classic still available from the Optical Society of America.
References
Aach, H.
1970
1972
On the use and phenomena of fluorescent pigments in paintings.
Leonardo 3:135-8.
Fluorescent pigments in art and design. j ournal of Color and Appearance
1(6) :25-28.
Agoston, G. A.
1987 Color Theory and Its Application in Art and Design, 2d ed. Berlin:
Allen, E.
1970
Springer.
An index of metamerism for observe differences. In Color 69, ed.
Manfred Richter. Giittingen, Germany: Musterschmidt.
1973 Prediction of optical properti es of paints from theory (with special refer-
ence to microvoid paints). journal of Paint Technology 45(584): 65- 72.
American Society for Testing and Materials (ASTM)
ASTM Standard D 156
1994 Standard rest method for Saybolt color of petroleum products (Saybolt
chromometer method). In Annual Book of ASTM Standards. Philadelphia:
American Society for Testing and Materials.
ASTM Standard D 523
1989 Standard test method for specular gloss. In Annual Book of ASTM
Standards. Philadelphia: American Society for Testing and Materials.
ASTM Standard D 985
1993 Standard test method for brightness of pulp, paper, and paperboard
(directional reflectance at 457 nm). In Annual Book of ASTM Standards.
Philadelphia: American Society for Testing and Materials.
ASTM Standard D 1003
1997 Standard test method for haze and lumi nous transmittance of transparent
plastics. In Annual Book of ASTM Standards. West Conshohocken, Pa.:
American Society for Testing and Materials.
ASTM Standard D 1209
1997 Standard test method for color of clear liquids (platinum-cobalt scale). In
Annual Book of ASTM Standards. West Conshohocken, Pa.: American
Society for Testing and Materials.
ASTM Standard D 1494
1992 Standard test method for diffuse light transmission factor of reinforced
plastics panels. In Annual Book of ASTM Standards. Philadelphia:
American Society for Testing and Materials.
References 333
ASTM Standard D 1500
1996 Standard test method for ASTM color of petroleum products (ASTM
color scale). In Annual Book of ASTM Standards. Philadelphia: American
Society for Testing and Materials.
ASTM Standard D 1535
1996 Standard practice for specifying color by the Munsell system. In Annual
Book of ASTM Standards. Philadelphia: American Society for Testing
and Materials.
ASTM Standard D 1544
1998 Standard test method for color of transparent liquids (Gardner color
scale). In Annual Book of ASTM Standards. West Conshohocken, Pa.:
American Society for Testing and Materials.
ASTM Standard D 1746
1997 Standard test method for transparency of plastic sheeting. In Annual
Book of ASTM Standards. West Conshohocken, Pa.: American Society
for Testing and Materials.
ASTM Standard D 1925
(Discontinued in 1995.) Standard test method for yellowness index of plastics. In
Annual Book of ASTM Standards. Philadelphia: American Society for
Testing and Materials.
ASTM Standard D 2244
1993 Standard test method for calculation of color differences from instrumen-
tally measured color coordinates. In Annual Book of ASTM Standards .
Philadelphia: American Society for Testing and Materials.
ASTM Standard D 2805
1996 Standard test method for hiding power of paints by reflectometry. In
Annual Book of ASTM Standards. Philadelphia: American Society for
Testing and Materials.
ASTM Standard D 4838
1988 Standard test method for determining the relative tinting strength of
chromatic paints. In Annual Book of ASTM Standards. Philadelphia:
American Society for Testing and Materials.
ASTM Standard E 259
1993 Standard practice for preparation of pressed powder white reflectance
factor transfer standards for hemispherical geometry. In Annual Book
of ASTM Standards. Philadelphia: American Society for Testing and
Materials.
ASTM Standard E 284
1997 Standard terminology of appearance. In Annual Book of ASTM
Standards. West Conshohocken, Pa.: American Society for Testing
and Materials.
ASTM Standard E 308
1996 Standard practice for computing the colors of objects by using the CIE
system. In Annual Book of ASTM Standards. Philadelphia: American
Society for Testing and Materials.
ASTM Standard E 313
1996 Standard practice for calculating yellowness and whiteness indices from
instrumentally measured color coordinates. In Annual Book of ASTM
Standards. Philadelphia: American Society for Testing and Materials.
334 References
ASTM Standard E 805
1994 Standard practice for identification of instrumental methods of color or
color-difference measurement of materials. In Annual Book of ASTM
Standards. Philadelphia: American Society for Testing and Materials.
ASTM Standard E 808
1998 Standard practice for describing retroreflection. In Annual Book of ASTM
Standards. West Conshohocken, Pa.: American Society for Testing and
Materials.
ASTM Standard E 809
1994 Standard practice for measuring photometric characteristics of
retroreflectors. In Annual Book of ASTM Standards. Philadelphia:
American Society for Testing and Materials.
ASTM Standard E 810
1994 Standard test method for coefficient of retroreflection of retroreflective
sheeting. In Annual Book of ASTM Standards. Philadelphia: American
Society for Testing and Materials.
ASTM Standard E 811
1995 Standard practice for measuring colorimetric characteristics of
retroreflectors under nighttime conditions. In Annual Book of ASTM
Standards. Philadelphia: American Society for Testing and Materials.
ASTM Standard E 911
1997 Standard practice for color measurement of fluorescent specimens. In
Annual Book of ASTM Standards. West Conshohocken, Pa.: American
Society for Testing and Materials.
ASTM Standard E 1164
1994 Standard practice for obtaining spectrophotometric data for object-color
evaluation. In Annual Book of ASTM Standards. Philadelphia: American
Society for Testing and Materials.
ASTM Standard E 1247
1992 Standard test method for identifying fluorescence in object-color speci-
mens by spectrophotometry. In Annual Book of ASTM Standards.
Philadelphia: American Society for Testing and Materials.
ASTM Standard E 1331
1996 Standard test method for reflectance factor and color by spectrophotome-
try using hemispherical geometry. In Annual Book of ASTM Standards.
Philadelphia: American Society for Testing and Materials.
ASTM Standard E 1336
1996 Standard test method for obtaining colorimetric data from a video display
unit by spectroradiometry. In Annual Book of ASTM Standards.
Philadelphia: American Society for Testing and Materials.
ASTM Standard E 1341
1996 Standard practice for obtaining spectroradiomerric data from radiant
sources for colorimetry. In Annual Book of ASTM Standards.
Philadelphia: American Society for Testing and Materials.
ASTM Standard E 1345
1992 Standard practice for reducing the effect of variability of color measure-
ment by use of multiple measurements. In Annual Book of ASTM
Standards. Philadelphia: American Society for Testing and Materials.
References 335
ASTM Standard E 1348
1990 Standard test method for transmittance and color by spectrophotometry
using hemispherical geometry. In Annual Book of ASTM Standards. West
Conshohocken, Pa.: American Society for Testing and Materials.
ASTM Standard E 1349
(Discontinued in 1997.) Standard test method for reflectance factor and color by
spectrophotometry using bidirectional geometry. In Annual Book of
ASTM Standards. West Conshohocken, Pa.: American Society for Testing
and Materials.
ASTM Standard E 1360
1990 Standard practice for specifying color by using the Optical Society
of America uniform color scales system. In Annual Book of ASTM
Standards. Philadelphia: American Society for Testing and Materials.
ASTM Standard E 1541
1994 Standard practice for specifying and matching color using the Colorcurve
system. In Annual Book of ASTM Standards. Philadelphia: American
Society for Testing and Materials.
ASTM Special Technical Publication 4 78
1970 Appearance of Metallic Surfaces. Philadelphia: American Society for
Testing and Materials. (Out of print. Authorized facsimile available
from University Microfilms International, 300 N. Zeeb Road, Ann Arbor,
Ml 48106.)
Andrews, C. L.
1960 Optics of the Electromagnetic Spectrum. Englewood Cliffs, N.J.:
Prentice Hall.
Armstrong, W. G., and W. D. Ross
1966 Anomalous reflectance from films pigmented with titanium dioxide.
Journal of Paint Technology 38(499):462-68.
Atkins, J. T., and F. W. Billmeyer Jr.
1966 Edge-loss errors in reflectance and transmittance measurement of translu-
cent materials. Materials Research and Standards 6:564-70.
1968 On the interaction of light with matter. Color Engineering 6(3):40.
Baer, N. S., A. Joel, R. L. Feller, and N. Indictor
1986 Indian Yellow. In Artists' Pigments: A Handbook of Their History and
Characteristics, vol. 1, ed. R. L. Feller, 17-36. Washington, D.C.:
National Gallery of Art.
Bartleson, C. J.
1974 Gloss measurements. Report to CIE Technical Committee 2.3. Vienna:
CIE Central Bureau.
Beer, A.
1852 Bestimmung der Absorption des rothen Lichts in farbigen Fliissizkeiten
(Determination of the absorption of red lights in colored liquids).
Annalen der physikalische Chemie 86(2):78-88.
Berger-Schunn, A.
1994 Practical Color Measurement. New York: John Wiley and Sons.
336 References
Berndt, H.
1989 A reexamination of paper yellowing and the Kubeika-Munk theory.
In Historic Textile and Paper Materials, ed. S. H. Zeronian and H. L.
Needles, vol. 2, 81- 91. ACS Symposium Series 410. Washington, D.C.:
American Chemical Society.
Billmeyer, F. W., Jr., and D. H. Alman
1973 Exact calculati on of Fr.esnel reflection coefficients for diffuse light.
j ournal of Color and Appearance 2(1):36-38.
Billmeyer, F. W., Jr., and D. G. Phillips
1974 Predicting reflectance and color of paint films by Kubeika-Munk analysis.
III. Effect of concentration errors on color for mixtures of one chromatic
pigment with white. j ournal of Paint Technology 46(592) :36- 39.
Billineyer, F. W. , Jr., and M. Saltzman
1981 Principles of Color Technology, 2d ed. New York: John Wiley and Sons.
(The third edition of this work appeared in 2000 under the title Billmeyer
and Saltzman's Principles of Color Technology, by Roy S. Berns [New
York: Wiley and Sons]).
Bouguer, P.
1729
Braun, J. H.
1991
Brockes, A.
1960
1964
Essai d' optique sur la gradation de la lumiere (Essay on the optics of
gradation of light ). Paris: Claude Tombert.
Gloss of paint films and the mechani sm of pigment involvement . journal
of Coatings Technology 63(799):43-51.
The influence of glossy surfaces on remission measurements. Farbe
9(13):53-62.
Der Zusammenhang von Farbstiirke und Teilchengrosse von
Buntpigmenten nach der Mie-theorie (The dependence of color strength
and particle size of colored pigments according to the Mie theory). Optik
21(10):550-56.
1969 Vergleich der Metamerie-Indizes bei Lichtartwechsel von Tageslicht zur
Gliihlampe und zu verschiedenen Leuchstofflampen (Comparison of
metameric indices wi th the change in the kind of light from daylight to
incandescent and to various fluorescent lamps). Farbe 18:233-39.
1970 Vergleich von berechneten Metamerie-Indi zes mit Abmusterungsergebissen
(Comparison of calculated metameric indices wi th the results of the
[visual] evaluations of samples) . Farbe 19:135-44.
Brown, W.R.]., and D. L. MacAdam
1957 Color discriminati ons of twelve observers. j ournal of the Optical Society
of America 47:137-43.
Bue, G. L., R. H. Kienle, L. A. Melsheimer, and E. I. Stearns
1947 Phenomenon of bronzing and surface coatings. Industrial and Engineering
Chemistry 39: 14 7- 54.
Burnham, R. W., R. M. Hanes, and C. ]. Bartleson
1963 Color: A Guide to Basic Facts and Concepts. New York: John Wiley
and Sons.
Burrell, H.
1971 Hiding without pigments. journal of Paint Technology 43(559):48- 53.
References
Buttignol, V.
1968 Optical behavior of iron oxide pigments. Journal of Paint Technology
40(526):480-93.
337
Byler, W. H.
1973 Luminescent pigments, inorganic. In Pigment Handbook, ed. T. C. Patton,
vol. 1, 905-23. New York: John Wiley and Sons.
Carr, W.
1973 Pigment powders and dispersions. In Pigment Handbook, ed. T. C.
Patton, vol. 3, 11-27. New York: John Wiley and Sons.
Celikiz, G., and R. G. Kuehni, eds.
1983 Color Technology in the Textile Industry. Research Triangle Park, N.C.:
Chase, T.
1982
American Association of Textile Chemists and Colorists.
Poster Session, Washington Congress, International Institute for
Conservation of Historic and Artistic Works, September 3-9.
Chickering, K. D.
1967 Optimization of the MacAdam modified 1965 Friele color difference
formula. journal of the Optical Society of America 57:537-41.
1971 FMC color-difference formulas: Clarification concerning usage. Journal of
the Optical Society of America 61:118-22.
Commission Internationale de l'Eclairage (CIE)
1972 Special metamerism index: Change of illuminant. Supplement 1 to
Colorimetry, CIE publication 15. Vienna: CIE Central Bureau.
1977 Radiometric and Photometric Characteristics of Materials and Their
Measurement. CIE publication 38. Vienna: CIE Central Bureau.
1982 Retroreflection: De'{inition and Measurement. CIE publication 54. Vienna:
CIE Central Bureau.
1984 The Spectroradiometric Measurement of Light Sources. CIE Publication
63. Vienna: CIE Central Bureau.
1986 Colorimetry, 2d ed. Official Recommendations of the International
Commission of Illumination. CIE publication 15.2. Vienna: CIE Central
Bureau.
1987 International Lighting Vocabulary, 4th ed. CIE publication 17.4. Vienna:
CIE Central Bureau.
1995 Method of Measuring and Specifying Colour Rendering Properties of
Light Sources, 3d ed. CIE publication 13.3. Vienna: CIE Central Bureau.
Cornman, M.
1986 Cobalt yellow (Aureolin). In Artists' Pigments: A Handbook of Their
History and Characteristics, vol. 1, ed. R. L. Feller, 37-46. Washington,
D.C.: National Gallery of Art.
Craker, W. E., and F. D. Robinson
1967 The effect of pigment volume concentration and film thickness on the
optical properties of surface coatings. journal of the Oil and Colour
Chemists Association 50: 111-33.
Davidson, H. R.
1978 Preparation of the OSA uniform color scales committee samples. Journal
of the Optical Society of America 68:1141-42.
338 References
Davidson, H. R. , and H. Hemmendinger
1966 Color prediction using the two-constant turbid-media theory. Journal of
the Optical Society of America 56:1102-9.
Davidson, H. R., H. Hemmendinger, and J. R. R. Landry
1963 A system of instrumental colour control for the texti le industry. Journal
of the Society of Dyers and Colourists 79:577- 89.
de la Rie, E. R.
1982 Fluorescence of paint and varnish layers, parts 1, 2, and 3. Studies in
Conservation 27:1-7; 65-79; 102-8.
Derby, R. E., Jr.
1952 Applied spectrophotometry. I. Color matching with the aid of the ' R'
Cam. American Dyestuff Reporter 41:550-57.
De Vore, J. R. , and A.H. Pfund
1947 Optical scattering by dielectric powders of uniform particle size. Journal
of the Optical Society of America 3 7:826-32.
DiBernardo, A., and P. Resnick
1959 Anomalous behavior of copper phthalocyanine-benzidine yellow mix-
tures. Journal of the Optical Society of America 49:480.
Duncan, D. R.
1962
Dunn, E. J.
1973
The identification and estimation of pigments in pigmented compositions
by reflectance spectrophotometry. Journal of the Oil and Colour Chemists
Association 45 :300-324.
White hiding lead pigments. In Pigment Handbook, ed. T. C. Patton, vol.
1, 65-84. New York: John Wiley and Sons.
Erhardt, 0., and J. S. Tsang
1990 The extractable components of oil paint films. In Cleaning, Retouching,
and Coatings: Preprints of the Contributions to the Brussels Congress,
ed. J. S. Mills and P. Smith, 93-97. London: International Institute for
Conservation of Historic and Artistic Works.
Evans, R. M.
1948 An Introduction to Color. New York: John Wiley and Sons.
1959 Eye, Film, and Camera in Color Photography. New York: John Wiley
and Sons.
1971 The perception of color. In Industrial Color Technology, ed. R. F. Gould,
43-68. Advances in Chemistry Series 107. Washington, D.C.: American
Chemical Society.
1974 The Perception of Color. New York: John Wiley and Sons.
Federation of Societies for Coatings Technology (FSCT)
1991 Infrared Spectroscopy Atlas for the Coatings Industry, 4th ed. Blue Bell,
Pa.: Federation of Societies for Coatings Technology.
1995 Coatings Encyclopedic Dictionary, ed. Stanley LeSota. Blue Bell, Pa.:
Feller, R. L.
1967
Federation of Societies for Coatings Technology.
Studies on the darkening of vermi li on by light. In Report and Studies on
the History of Art, 99-11 1. Washington, D.C.: National Gallery of Art.
References 339
1968 Problems in reflectance spectrophotometry. In IIC-1967 London
Conference on Museum Climatology, 257-69. London: International
Institute for Conservation of Historic and Artistic Works.
Feller, R. L., ed.
1986 Artists' Pigments: A Handbook of Their History and Characteristics,
vol. 1. Washington, D.C. : National Gallery of Art (softcover); Cambridge:
Cambridge University Press (hardcover).
Feller, R. L., M. Curran, and C. W. Bailie
1984 Identification of traditional organic colorants employed in Japanese prints
and determination of their rates of fading. In Japanese Woodblock Prints:
A Catalogue of the Mary A. Ainsworth Collection. Oberlin, Ohio: Allen
Memorial Art Museum.
Feller, R. L., and R. M. Johnston-Feller
1997 Vandyke Brown, Cologne and Kassel earth. In Artists' Pigments, A
Handbook of Their History and Characteristics, vol. 3., ed. Elisabeth
West FitzHugh. Washington, D.C.: National Gallery of Art.
Feynman, R. P.
1985 QED, the Strange Theory of Light and Matter. Princeton, N.J.: Princeton
University Press.
Fiedler, I., and M. Bayard
1986 Cadmium yellows, oranges, and reds. In Artists' Pigments: A Handbook
of Their History and Characteristics, vol. 1, ed. R. L. Feller, 65-108.
Washington, D.C.: National Gallery of Art.
Fraser, I.
1973 Centrifugal sedimentation. In Pigment Handbook, ed. T. C. Patton,
vol. 3, 53-62. New York: John Wiley and Sons.
Fuller, C. W.
1973 Colored iron oxide pigments, synthetic. In Pigment Handbook, ed. T. C.
Patton, vol. 1, 333-49. New York: John Wiley and Sons.
Fuller, N. A.
1985 Analysis of thin-layer chromatograms of paint pigments and dyes
by direct microspectrophotometry. Forensic Science International
27:189-204.
Garland, C. E.
1983 A general procedure for determination of relative dye strength by spectro-
photometric measurement of reflectance factor. In Color Technology in
the Textile Industry, ed. G. Celikiz and R. G. Kuehni, 107-12. Triangle
Park, N.C.: American Association of Textile Chemists and Colorists.
Garrett, M. D.
1973 Carbon black, bone black, lamp black. In Pigment Handbook, ed. T. C.
Patton, vol. 1, 709-43. New York: John Wiley and Sons.
Gerstner, W.
1966 Crystal form and particle size of organic pigments in printing inks and
paints. Journal of the Oil and Colour Chemists Association 49(11):954.
Also reprinted by Badische Anilin+Soda-Fabrik AG, Ludwigshafen am
Rhein, Germany.
Gettens, R. J., R. L. Feller, and W. T. Chase
1993 Vermilion and cinnabar. In Artists' Pigments: A Handbook of Their
History and Characteristics, vol. 2, ed. A. Roy, 159-82. New York:
Oxford University Press.
340 References
Gettens, R. J., and E.W. FitzHugh
1993a Azurite and blue verditer. In Artists' Pigments: A Handbook of Their
History and Characteristics, vol. 2, ed. A. Roy, 23-36. New York: Oxford
University Press.
1993b Malachite and green verditer. In Artists' Pigments: A Handbook of Their
History and Characteristics, vol. 2, ed. A. Roy, 183-202. New York:
Oxford University Press.
Gettens, R. J. , and G. L. Stout
1942 Painting Materials. A Short Encyclopedia. New York: D. Van Nostrand
Co. (Reprint, New York: Dover, 1966.)
Giovanelli, R. G.
1956 A note on the reflection of internally incident di ffuse light. Optica Acta
3: 127.
Granville, W. C.
1994 The color harmony manual, a color atlas based on the Ostwald system.
Color Research and Application 19:77-98.
Greenstein, L. M.
1973a Nacreous (pearlescent ) pigments. In Pigment Handbook, ed. T. C. Patton,
vol. 1, 871-90. New York: John Wiley and Sons.
1973b Pearlescence: The optical behavior of nacreous and interference pigments.
In Pigment Handbook, ed. T. C. Patton, vol. 3, 357-90. New York: John
Wiley and Sons.
1988 Nacreous (pearl escent pigments) and interference pigments. Mearl
Technical Bulletin Number 2. New York: Mearl Corp. (This is a corrected
version of the chapter in the 2d ed. of Pigment Handbook, ed. T. C.
Patton, vol. 1 [New York: John Wiley and Sons].)
Grissom, C. A.
1986 Green Earth. In Artists' Pigments: A Handbook of Their History and
Characteristics, vol. 1, ed. R. L. Feller, 141-68. Washington, D.C.:
National Gallery of Art.
Grum, F.
1980 Colorimetry of fluorescent materials. In Optical Radiation Measurement.
Vol. 2, Color Measurement, ed. F. Grum and C. J. Bartleson, 241-88.
New York: Academic Press.
Hansen, E. F., R. Lowinger, and E. Sadoff
1993 Consolidation of porous paint in a vapor-saturated atmosphere: A tech-
nique for minimizing changes in the appearance of powdering matte
paint. j ournal of the American Institute for Conservation 32(1):1-14.
Hansen, E. F., S. Walston, and M. H. Bishop, eds.
1993 Matte paint. Bibliographic supplement to Art and Archeology Technical
Abstracts 30.
Hard, A., and L. Sivik
1981 NCS-narural color system: A Swedish standard for color notation. Color
Research and Application 6:129-38.
Hardy, A. C.
1936 Handbook of Colorimetry. Cambridge, Mass.: Technology Press
(Massachusetts Institute of Technology).
References
Hay, G. N.
1973 Stai nless steel flake pigment. In Pigment Handbook, ed. T. C. Patton,
vol. 1, 833-35. New York: John Wiley and Sons.
Hemmendinger, H.
341
1983 The calibration of a spectrophotometer for color measurement. In Color
Technology in the Textile Industry, ed. G. Celikiz and R. G. Kuehni,
35-48. Research Triangle Park, N.C.: American Association of Textile
Chemists and Colorists.
Hemmendinger, H., and R. M. Johnston
1970 A goniospectrophotometer for color measurements. In Proceedings,
AIC Color 69, ed. M. Richter, vol. 1, 509. Gottingen, Germany:
Musterschmidt.
Hislop, R. W., and P. L. McGinley
1978 Microvoid coatings: Pigmented, vesiculated beads in flat latex paints.
Journal of Coatings Technology 50(642):69-77.
Huebner, F. E., and H. N. Monck
1992 Measurement of color in resins and adhesive systems. Tappi Journal
75(9): 197-204.
Hunter, R. S.
1938 Methods of determining gloss. Journal of Research, National Bureau of
Standards 18:19.
1942 Photoelectric tristimulus colorimetry with three filters. National Bureau
of Standards Circular C429, July 30. Washington, D.C.: U.S. Government
Printing Office. Reprinted in Journal of the Optical Society of America
32:509-38.
1958 Photoelectric color difference meter. Journal of the Optical Society of
America 48:985-95.
Hunter, R. S., and D. B. Judd
1939 Development of a method of classifying paints according to gloss.
American Society for Testing and Materials Bulletin 97:11-18.
Illuminating Engineering Society (JES)
1983 Guide to Spectroradiometric Measurements. New York: Illuminating
Engineering Society (American National Standards Institute).
International Electrochemical Commission (IEC)
1974 Photometric and Colorimetric Methods of Measurement of the Light
Emitted by a Cathode-Ray Tube Screen. Geneva: International
Electrochemical Commission.
Jacobsen, A. E.
1948 Non-adaptability of the ICI system to some near-whites which show
absorption in the far-blue region of the spectrum. Journal of the Optical
Society of America 38:442.
Jacobson, E.
1948 Basic Color: An Interpretation of the Ostwald Color System. Chicago:
Paul Theobald.
Johnston, R. M.
1963 Pitfalls in color specifications. Official Digest, Federation of Societies for
Paint Technology 35:259-74.
1967a Geometric meramerism. Color Engineering 5(3):42-47, 54.
342 References
1967b
1969
1971a
1971b
1973
Spectrophotometry for the analysis and description of color. journal of
Paint Technology 39(509):346-54.
Color control in the small paint plant. Journal of Paint Technology
41 (534):415-21.
Color measuring instruments: A guide to their selection (a report of Inter-
Society Color Council). journal of Color and Appearance 1(2):27-30.
Reprinted in 1983 with additional prefatory remarks in Color Technology
in the Textile Industry, ed. G. Celikiz and R. G. Kuehni, 21-33. Research
Triangle Par k, N.C.: American Association of Textile Chemists and
Colorists.
Colorimetry of transparent materials. journal of Paint Technology
43(553): 42-50.
Color theory. In Pigment Handbook, ed. T. C. Patton, vol. 3, 229-88.
New York: John Wiley and Sons.
Johnston, R. M., and R. L. Feller
1963 The use of differential spectral curve analysis in the study of museum
objects. Dyestuffs 44(9): 1-10.
1967 Optics of paint films : Glazes and chalking. In Application of Science in
Examination of Works of Art, 86-95. Boston: Museum of Fine Arts.
Johnston, R. M. , and R. Stanziola
1969 Angular Color Measurement on Automotive Materials. Preprint no.
690241. Detroit: International Automotive Engineering Conference.
Johnston-Feller, R. M.
1986 Reflections on the phenomenon of fading. journal of Coatings Technology
58(736):32-50.
Johnston-Feller, R. M., and C. W. Bailie
1982a An analysis of the optics of paint glazes: Fading. In Science and
Technology in the Service of Conservation, 180-85. Preprints of contribu-
tions to the Washington Congress. London: International Institute for
Conservation of Historic and Artistic Works.
1982b Determination of the tinting strength of chromatic pigments. Journal of
Coatings Technology 54(692):43-56.
Johnston-Feller, R. M., R. L. Feller, C. W. Bailie, and M. Curran
1984 The kinetics of fading: Opaque paint films pigmented with alizarin lake
and titanium dioxide. Journal of the American Institute for Conservation
23:114-29.
Johnston-Feller, R. M., and D. Osmer
1977 Exposure evaluation: Quantification of changes in appearance of pig-
mented materials. journal of Coatings Technology 49(625):25-36.
1979 Exposure evaluation, part 2-Bronzing. Journal of Coatings Technology
51(650):37-44.
Jordan, R. D., and M.A. O'Neill
1991 The whiteness of paper. Colorimetry and visual ranking. Tappi Journal
74(5):93-101.
Judd, D. B.
1937 Optical specification of light-scattering materials. journal of Research,
National Bureau of Standards 19:287-31 7.
References
1939 Specification of color tolerances at the National Bureau of Standards.
American journal of Psychology 52:418-28.
343
1942 Fresnel reflection of diffusely incident light. Journal of Research, National
Bureau of Standards 29:329-32.
1950 Colorimetry. National Bureau of Standards Circular 4 78. Washington,
D.C. : U.S. Government Printing Office.
Judd, D. B., and G. Wyszecki
1963 Color in Business, Science, and Industry, 2d ed. New York: John Wiley
and Sons.
1975 Color in Business, Science, and Industry, 3d ed. New York: John Wiley
and Sons.
Kampfer, W. A.
1973 Titanium dioxide. In Pigment Handbook , ed. T. C. Patton, vol. 1, 1-36.
New York: John Wiley and Sons.
Kelly, K. L., and D. B. Judd
1955 The ISCC-NBS Method of Designating Colors and a Dictionary of Color
Names. National Bureau of Standards Circular 553. Washington, D.C.:
U.S. Government Printing Office (out of print}.
1976 Color: Universal Language and Dictionary of Names. National Bureau of
Standards Special Publication 440. Washington, D.C.: U. S. Government
Printing Office. (Available as a photocopy in black and white from the
National Technical Information Service, Springfield, VA 221i'l. Order
number PB265225.)
Kerker, M., and D. D. Cooke
1976 Pigmented microvoid coatings II. Luminance and color of the concentric
sphere model. journal of Coatings Technology 48(621):35-41.
Kerker, M., D. D. Cooke, and W. D. Ross
1975 Pigmented microvoid coatings. Theoretical study of three models. Journal
of Paint Technology 47(603):33-42.
Kettler, W.
1995a
1995b
Kortiim, G.
Color measuring instruments. Part 1: Portable systems. European
Coatings Journal 1 and 2:210-12, 214, 216.
Color measuring instruments. Part 2: Stationary systems. European
Coatings Journal 3:59-62.
1969 Reflectance Spectroscopy: Principles, Methods, Applications. Translated
from German by James E. Lohr. New York: Springer-Verlag.
Kremer Pigments, Inc.
1992 61 East 3d Street, New York, NY 10003. Supplier of artists' pigments,
paints, and vehicles; a branch of Dr. George F. Kremer, Farbmiihle,
D-W-7974 Aichsteiten, Germany.
Kubeika, P.
1948
1954
New contributions to the optics of intensely light scattering materials,
part 1. Journal of the Optical Society of America 38:448-57, 1067.
New contributions to the optics of intensely light scattering materials,
part II. Non-homogeneous layers. journal of the Optical Society of
America 44:330-35.
344 References
Kubeika, P., and F. Munk
1931 Ein Beitrag zur Optik der Farbanstriche. Zeitschrift fur technische Physik
12:593-601.
Kuehni, R.
1975
1983
Kiihn, H.
Computer Colorant Formulation. Lexington, Mass.: D.C. Heath and
Company.
A general procedure for the determination of relative dye strength by
spectrophotometric transmittance measurement. In Color Technology in
the Textile Industry, ed. G. Celikiz and R. G. Kuehni, 77-89. Research
Triangle Park, N.C.: American Association of Textile Chemists and
Colorists.
1993a Lead-tin yellow. In Artists' Pigments: A Handbook of Their History and
Characteristics, vol. 2, ed. A. Roy, 83-112. New York: Oxford University
Press.
1993b
Kumar, R.
Verdigris and copper resinate. In Artists' Pigments: A Handbook of Their
History and Characteristics, vol. 2, ed. A. Roy, 131-58. New York:
Oxford University Press.
1981 Analysis for organic pigments using solution spectrophotometry. Ph.D.
diss., Rensselaer Pol ytechnic Institute.
Lafontaine, R. H.
1986 Seeing through a yel low varnish: A compensating illumination system.
Studies in Conservation 31:97-102.
Lambert, J. H.
1760 Photometria sine de Mensura et Gradibus Luminis Co/arum et Umbrae
(Photometry concerning the measurement and gradation of colored light
and shadows). Augsburg, Germany: Eberhard Klett.
LaMer, V. K., J. Q. Umberger, F. Sinclair, and F. E. Buchwalter
1945 OSRD Report 4904, PB32208. U.S. Department of Commerce, Office of
Scientific Research and Development.
Laurie, A. P.
1926 The Painter's Methods and Materials. London: Seeley Service.
Lettieri, T. R., E. Marx, J.-F. Son, and T. V. Borburger
1991 Light scattering from glossy coatings on paper. Applied Optics
30:4439-47.
Levison, H. W.
1976 Artists' Pigments: Lightfastness Tests and Ratings. Hallandale, Fla.:
Love, C.H.
1973
Colorlab.
Colored iron oxide pigments, natural. In Pigment Handbook, ed. T. C.
Patton, vol. 1, 323-32. New York: John Wiley and Sons.
MacAdam, D. L.
1942 Visual sensitivities to color differences in daylight. Journal of the Optical
Society of America 32:247- 74.
1943 Specification of small chromaticity differences. Journal of the Optical
Society of America 33: 18-26.
1974 Uniform color scales. Journal of the Optical Society of America
64:1619-1702.
References 345
1978 Colorimetric data for samples of OSA uniform color scales. journal of the
Optical Society of America 68:121-30.
McCarley, ]. E., C. E. Green, and K. H. Horowitz
1965 Digital system for converting spectrophotometric data to CIE coordinates,
dominant wavelength, and excitation purity. journal of the Optical
Society of America 55 :355-60.
McCrone, W. C.
1973a Size, shape, and size distribution of pigments. In Pigment Handbook,
ed. T. C. Patton, vol. 3, 63-69. New York: John Wiley and Sons.
1973b Pigment identification, microscopy. In Pigment Handbook, ed. T. C.
Patton, vol. 3, 71-76. New York: John Wiley and Sons.
McDonald, R.
1980 Industrial pass-fail colour matching. Part I-Preparation of visual colour-
matching data; Part II-Methods of fitting tolerance ellipsoids; Part III-
Development of pass-fail formula for use with instrumental measurement
of colour difference. journal of the Society of Dyers and Colourists
96:372-76; 418-33; 486-95.
McLaren, K.
1986 The Colour Science of Dyes and Pigments, 2d ed. Bristol, Eng.:
Adam Hilger.
Maerz, A., and M. R. Paul
1930 Dictionary of Color. New York: McGraw-Hill Book Co.
Maxwell, J. C.
1860 On the theory of compound colours and the relations of the colours of
the spectrum. Proceedings of the Royal Society of London 10:404-9, 484.
Michalski, S.
1990 A physical model of the cleaning of oil paint. In Cleaning, Retouching,
and Coatings: Preprints of the Contributions to the Brussels Congress, ed.
J. S. Mills and P. Smith. London: International Institute for Conservation
of Historic and Artistic Works.
Mie, G.
1908 Beitrage zur Optik tri.iber Medien, speziell kolloidalen Metallosungen
(Contributions to the optics of turbid media, especially colloidal metal
solutions). Annalen der Physik 25:377-445. (A brief presentation in
English was made by J. A. Stratton and H. D. Houghton, 1931,
Theoretical investigation of light through fog. Physical Review
38:159-65.)
Mitton, P. B.
1973 Opacity, hiding power, and tinting strength. In Pigment Handbook,
ed. T. C. Patton, vol. 3, 289-329. New York: John Wiley and Sons.
Mudgett, P. S., and L. W. Richards
1973 Kubeika-Munk scattering and absorption coefficients for use with glossy
opaque objects. Journal of Paint Technology 45(586):43-53.
Miihlethaler, B., and J. Thissen
1993 Smalt. In Artists' Pigments: A Handbook of Their History and
Characteristics, vol. 2, ed. A. Roy, 113-30. New York: Oxford University
Press.
Munsell, A. H.
1905 A Color Notation. Boston: Ellis.
346
References
Munsell Book of Color
Available only from GretagMacbeth, Inc., Munsell Color Dept., 617 Little Britain
Road, New Windsor, NY 12553-6148.
Nassau, K.
1983 The Physics and Chemistry of Color. New York: John Wiley and Sons.
National Bureau of Standards
1968-71 Fundamental principles of absolute radiometry and the philosophy of the
NBS program. In National Bureau of Standards Technical Note 594-1.
Washington, D. C. : U.S. Government Printing Office.
Newhall, S. M., D. Nickerson, and D. B. Judd
1943 Final report of the O.S.A. subcommittee on the spacing of the Munsell
colors. j ournal of the Optical Society of America 33:385-418.
Nickerson, D.
1940 History of the Munsell color system and its scientific application. journal
of the Optical Society of America 30:575-80. Reprinted 1976 in Color
Research and Application 1:69-77.
1975
1977
1978
1981
Nimeroff, I.
1957
1966
Uniform color scales: Munsell conversion of OSA committee selection.
journal of the Optical Society of America 65:205-7.
History of the OSA committee on uniform color scales. Optics News
3(1):8-17.
Munsell renotations for samples of OSA uniform color scales. journal of
the Optical Society of America 68:1343-47.
OSA uniform color scales samples-a unique set. Color Research and
Application 6:7-33.
Propagation of errors in tristimulus colorimetry. journal of the Optical
Society of America 4 7:697-702.
Comparison of uncertainty ellipses calculated from two spectrophoto-
metric colorimetry methods by an automatic-computer program. journal
of the Optical Society of America 56:230- 37.
Nimeroff, I., and J. A. Yurow
1965 Degree of metamerism. Journal of the Optical Society of America
55:185-90.
Novak, C. L., and S. A. Shafer
1992 Color vision. In The Encyclopedia of Artificial Intelligence, 192- 202.
New York: John Wiley and Sons.
Ogilvy, J. A.
1991 Theory of Wave Scattering from Random Rough Surfaces. Bristol, Eng.:
Adam Hilger.
Orchard, S. E.
1968 A new look at pigment optics. Journal of the Oil and Colour Chemists
Association 51:44-60.
Orr, C., Jr.
1973 Characterization of pigments: Gravity sedimentation techniques. In
Pigment Handbook, ed. T. C. Patton, vol. 3, 43-51. New York: John
Wiley and Sons.
References
Osmer, D.
1978
1982
The opacity parameter in color matching. Plastics Engineering 34:33.
Practical application of Kubeika-Munk theory for the near infrared.
Journal of Coatings Technology 54(692):35-41.
347
Ostwald, W.
1931 Colour Science. Authorized translation with an introduction by J. Scott
Taylor. Part I: Colour theory and colour standardization. Part II: Applied
colour science (1933). London: Winsor and Newton Ltd.
1969 The Colour Primer. Translated and edited, and with a foreword and
evaluation by Faber Birren. New York: Van Nostrand-Rheinhold.
Patton, T. C., ed.
1973 Pigment Handbook, vols. 1-3. New York: John Wiley and Sons.
Phillips, D. G., and F. W. Billmeyer Jr.
1976 Predicting reflectance and color of paint films by Kubeika-Munk analysis.
IV. Kubeika-Munk scattering coefficient. journal of Coatings Technology
48(616):30-36.
Pierce, P. E., and R. T. Marcus
1994 Color and Appearance. Blue Bell, Pa.: Federation of Societies for Coatings
Technology.
Pl esters, J.
1993 Ultramarine blue, natural and artificial. In Artists' Pigments: A Handbook
of Their History and Characteristics, vol. 2, ed. A. Roy, 37-54. New
York: Oxford University Press.
Raggi, A., and G. Barbiroli
1993 Colour difference measurement: The sensitivity of various instruments
compared. Color Research and Application 18:11-27.
Rheinholdt, W. C., and J. P. Menard
1960 Mechanized conversion of colorimetric data to Munsell renotation.
Journal of the Optical Society of America 50:802-7.
Richards, L. W.
1970 The calculation of the optical performance of paint films. journal of Paint
Technology 42:276-86.
Richter, M.
1953 Das System der DIN-Farbenkarte. Farbe 1:85-98.
1955 The official German standard color chart. journal of the Optical Society
of America 45:223-26.
Ridgway, R.
1912 Color Standards and Color Nomenclature. Washington, D.C.: Published
by author.
Rogers, P. E., F. S. Greenawald, and W. L. Butters
1973 Copper and copper alloy flake powders. In Pigment Handbook, ed. T. C.
Rolles, R.
1973
Patton, vol. 1, 807-17. New York: John Wiley and Sons.
Aluminum flake pigment. In Pigment Handbook, ed. T. C. Patton, vol. 1,
785-806. New York: John Wiley and Sons.
348 References
Rosenthal, W. S., and B. N. McBane
1973 Microvoids as light scattering sites in polymer coatings. Journal of Paint
Technology 45(584):73- 80.
Ross, W. D.
1967 Kubeika-Munk formulas adapted for better computation. Journal of
Paint Technology 39:515-21.
Saltzman, M.
1986 Analysis of dyes in museum textiles. In Textile Conservation Symposium
in Honor of Pat Reeves, ed. C. C. McLean and P. Connell, 27-39. Los
Angeles: Los Angeles County Museum of Art (Conservation Center ).
Saltzman, M., and A. M. Keay
1967 Colorant identification. Journal of Paint Technology 39(509):360-67.
Saunders, D.
1988 Color change measurement by digital image processing. National Gallery
Technical Bulletin (London) 12:66-77.
Saunders, D., and J. Cupitt
1993 Image processing at the National Gallery: The VASAR project. National
Gallery Technical Bulletin (London) 14:72-85.
Saunderson, J. L.
1942 Calculation of the color of pigmented plastics. Journal of the Optical
Society of America 32:727-36.
Schafer, H., and G. Wallisch
1981 Obtaining opacity with organic pigments in paint. Journal of the Oil and
Colour Chemists Association 64:405-14.
Schweppe, H., and H. Roosen-Runge
1986 Carmine. In Artists' Pigments: A Handbook of Their History and Charac-
teristics, vol. 1, ed. R. L. Feller, 255-84. Washington, D.C.: National
Gallery of Art.
Seiner, J. A.
1970
Seve, R.
1993
Opaque Films. British Patent 1,178,612. Assigned to PPG Industries, Inc.
Problems connected with the concept of gloss. Color Research and
Application 18:240-52.
Shurcliff, W. A.
1942 Curve shape index for identification by means of spectrophotometric
curves. Journal of the Optical Society of America 32:160-63.
Simon, H.
1971 The Splendor of Iridescence. New York: Dodd, Mead and Co.
Smith, D.
1962 Bronze phenomena in pigments and dyes. In RBH Trends, Interchemical
Corporation, part 1.
1963 Bronze phenomena in pigments and dyes. In RBH Trends, Interchemical
Corporation, part 2.
1992 The Catalog of Artists' Materials. Seattle: Daniel Smith.
Spindel, S.
1973 Coatings for copper. Paint and Varnish Production 63(2):21-27.
References
349
Staniforth, S.
1985 Retouching and colour matching: The restorer and metamerism. Studies
in Conservation 30:101-11.
Stanziola, R.
1980 Color differences caused by dye weighing errors. Color Research and
Application 5:129-32.
1992 The Colorcurve system. Color Research and Application 17:262-72.
Stearns, E. I.
1969 The Practice of Absorption Spectrophotometry. New York: John Wiley
and Sons.
Stenius, A. S.
1951 The application of the Kubeika-Munk theory to the diffuse reflection of
light from paper, parts 1 and 2. Svensk Papperstidning 54:663-70; 701.
1953 The application of the Kubeika-Munk theory to the diffuse reflection of
light from paper, part 3. Surface reflection and light scattering of paper.
Svensk Papperstidning 56:607-14.
Stiles, W. S., and G. Wyszecki
1968 Intersections of the spectral reflectance curves of metameric object colors.
Journal of the Optical Society of America 58:32-40.
Switzer, J. L., and R. C. Switzer
1950 Daylight Fluorescent Pigment Compositions. U.S. Patent 2,498,592.
Talsky, G., and M. Ristie-Solajie
1989 Higher order derivative reflectance spectrophotometry of synthetic
organic pigments in artists' paints. Analytica Chimica Acta 226:293-304.
Technical Association of the Pulp and Paper Industry (TAPP!)
1992a Method T260 om-91. Test to evaluate the aging properties of bleached
chemical pulps. In Tappi Test Methods. Atlanta: Technical Association of
the Pulp and Paper Industry.
1992b
1992c
1992d
Method T425 om-91. Opacity of paper (15 geometry/diffuse Illuminant
A, 89% reflectance backing). In Tappi Test Methods. Atlanta: Technical
Association of the Pulp and Paper Industry.
Method T452 om-92. Brightness of pulp, paper, and paperboard (direc-
tional reflectance at 457 nm). In Tappi Test Methods. Atlanta: Technical
Association of the Pulp and Paper Industry.
Method T519 om-91. Diffuse opacity of paper (D/0 paper backing).
In Tappi Test Methods. Atlanta: Technical Association of the Pulp and
Paper Industry.
Vagias, A. M.
1982 Thermoplastic Key Face for Pianos or the Like. U.S. Patent 4,346,639.
Van den Akker, J. A.
1949 Scattering and absorption of light in paper and other diffusing media. A
note on the coefficients of the Kubeika-Munk theory. Tappi 32:498-500.
Voedisch, R. W.
1973a Luminescent pigments, organic. In Pigment Handbook, ed. T. C. Patton,
vol. 1, 891-92. New York: John Wiley and Sons.
1973b Pigmentation of fluorescent paints. In Pigment Handbook, ed. T. C.
Patton, vol. 2, 143-49. New York: John Wiley and Sons.
350
References
Wainwright, I. N., J.M. Taylor, and R. D. Harley
1986 Lead antimonate yellow. In Artists' Pigments: A Handbook of Their
History and Characteristics, vol. 1, ed. R. L. Feller, 219-54. Washington,
D.C.: National Gallery of Art.
Walsh, J. W. T.
1926 The reflection factor of a polished glass surface for diffuse light.
Appendix of A. K. Taylor and C. ]. W. Grieveson, The Transmission
Factor of Commercial Window Glasses. Department of Scientific and
Industrial Research, Illumination Research Technical Paper No. 2.
London: Her Majesty's Stationery Office.
Weingrad, C.
1992 Three-dimensional color: Unlocking the mystery of interference pigments.
In Inksmith, spring: 9, 11-13. Daniel Smith Catalog of Artists' Materials.
Seattle: Daniel Smith. (Color reprints available for a small fee.)
Wyszecki, G.
1963 Proposal for a new color difference formula. Journal of the Optical
Society of America 53:1318-19.
Wyszecki, G., and W. S. Stiles
1982 Color Science: Concepts and Methods, Quantitative Data, and Formulae,
2d ed. New York: John Wiley and Sons.
Note: Letters (and t following page num-
bers indicate figures and tables, respec-
tively. Letters CP indicate color plates.
Aach, Herbert, 215-16, CPS
AATCC. See American Association of
Textile Chemists and Colorists
Absorbance, definition of, 115, 149
Absorption
changes in, from vehicle or substrate
yellowing, 102
as colorant characteristic, 58-61
definition of, 115
of glazes, 109-10
importance of, 14
of inorganic materials, 67
and tinting strength calculations,
128-30
particle size and, 133-39
for pigment mixtures, 130-33
in ultraviolet and infrared, 255-56
Absorption bands
organization of curves according to,
298,299-300
record of, suggested protocol for,
253t, 254
Absorption coefficient (A)
in Beer-Bouguer transmittance
equation, 64-66, 150
of organic materials, 66
of transparent materials, 64-65,
149-50. See also Optical density
Absorption coefficient (K), in Kubelka-
Munk equation, 67-70, 88-92
Absorption K/S curves, on logarithmic
scale
curve shapes, 147--49, 148(-154(
in exposure-related color changes,
98-100, 100(
for glazes, 112, 113(
Absorption maximum, in tinting strength
calculations, 128-30, 139
Absorption-scattering relationship,
reflectance and, 67-70, 68(
achromatic colorants, 70-72
near neutrals, 77
primary colorants, 72-74
Index
quantitative determination of K/S
values, 80-81
secondary colorants, 74-76
as single constant (K/S) in Kubelka-
Munk calculations, 80-86
Additive color mixtures, 16-17, 16(
color-matching functions for, 22-23, 23(
complementary, on chromaticity dia-
gram, 26-27
grayness of, 17, l 7(
Maxwell disk for, 17-18
measurement of, 20
primary colors for, 18, 19(
Additivity, chromaticity diagram for,
26,28(
Adirondack Red. See Red 2B-strontium
Aigami, 227
Air-filled voids. See Microvoids
Air-sample interface. See also Paint-to-air
interface
reflection at, 7
in Kubeika-Munk equation, 79
Alizarin
extender pigments in, 222
in natural madder lake, 207
reflectance curve of, 137, 137(
Alizarin crimson, inert pigment in, 223
Alizarin lake
fading of, 94, 95(, CPl-2
color-matching for, 104-5
as glaze
absorption K/S curve for, 112, 113(
critical concentration for maximum
chroma of, 114, CP3
reflectance curves for, 95(, 227, 228(
on inorganic carrier, 137, 137(
Alizarin maroon, reflectance curve for,
233,234(
Alumina, as inert pigment, 223
Aluminum flake, pigment
concentration of, 181
orientation of, 181, 181(
particle size of, 180
reflectance curves for, 182(-18 8(,
184-88
Aluminum hydrate, as extender pigment,
222
American Association of Textile Chemists
and Colorists (AATCC), 154
American Society for Testing and
Materials (ASTM), 154-55
Analog curve, from spectrophotometer, 6-7
Anatase Ti0
2
See Titanium white
Angle of illumination, 9-10, 10(
and metamerism, 41, 47--48
Angle of view, 10, 10(
and metamerism, 41, 47--48
ANLAB 40 equation, 34
Anthrapyrimidine yellow, reflectance curve
for, 230(
Anthraquinoid red
absorption band number and wave-
length of, 300t
reflectance curve for, 32 l(
Appearance
difference in, 115-16
interdisciplinary nature of, 258
Aromatic groups or solvents, in paint
vehicle, phthalo blue contamina-
tion in, 235
Artists' Pigments: A Handbook of Their
History and Characteristics
(Feller 1986), 53-54, 227-28,
229,233
Artists' Pigments: Light(astness Tests and
Ratings (Levison 1976), 222
Aspecular angle, 10, lOf
ASTM. See American Society for Testing
and Materials
Aureolin, reflectance curve for, 229-30
Averages, of measured data, 238, 240
Azo red, bronzed sample of
chromaticity coordinates of, 166, 167(
specular reflectance measurements of,
164-67, l67t
Azomethines, 229
Azurite, reflectance curve for, 227
Backscattering, 59
Barium sulfate (BaS0
4
). See also Blanc
fixe
as inert pigment, 223
pressed, 10, 11, 12, 194
refractive index of, 59
352
Beads
solid, 221
vesiculated, 221
Beer-Bouguer equation, 62, 66
for tinting strength of dyes, 128-29
for transparent materials, 62,
64-67, 150
Beer's law, 64, 65
Benzidine yellow
and phthalo blue, interaction between,
234-35,236(
reflectance curve for, 230(
Bidirectional geometry, in spectropho-
tometers, 8
for fluorescent sample measurements,
vs. integrating sphere, 207
Bismuth oxychloride (BiOCL), 172
Black
added to blue, reflectance curves for,
72-73
added to green, reflectance curves for,
75
added to purple, reflectance curves for,
75
added to red, reflectance curves for,
73-74
added to yellow, reflectance curves
for, 74
addition of
reflectance and, 68
saturation and, 14
mixtures
chromaticity diagram for, 268(
CIE notations for, 263t
reflectance curves for, 264(, 265(
Munsell notation for, 263t
in mixture with Ti0
2
, reflectance
curves for, 293(
reflectance curves for, 70- 72, 89- 91
Black port, in measurement of color of
metals, 8
Blanc fixe (BaS0
4
), 71, 223. See also
Barium sulfate
Bleaching, of vehicle or substrate, 102
and absorption K/S curve, 100, 100(
Bloom, 203-4
Blue, reflectance curves for, 72- 73,
225-27,226(
Blue-violet fluorescent dyes, as FWAs, 218
BON maroon (manganese)
absorption band number and wave-
length of, 300t
reflectance curve for, 308(
BON red
absorption band number and wave-
length of, 300t
chromaticity diagram of, 262(
CIE notation for, 273t
Munsell notation for, 273t
in neutral mixtures
CIE and Munsell notation for, 293t
reflectance curves for, 293(
in orange mixtures
CIE and Munsell notation for, 290t
reflectance curves for, 292(
Index
in purple mixtures
CIE and Munsell notation
for, 287t
reflectance curves for, 289(
reflectance curves for, 73- 74, 227,
228(, 276(
Bouguer's (Lambert's) law, 62-64
Brightness, of paper, industry standards,
155-57
Broadband filters, for tinting strength
measurements, 140-41
Brockes's curve, of absorption and scatter-
ing vs. particle size, 135, 136(
Brominated anthanthrone orange
absorption band number and wave-
length of, 300t
reflectance curves for, 325(
Brominated pyranthrone orange
absorption band number and wave-
length of, 300t
reflectance curves for, 324(
Bronze
color of, composition and, 160-62,
16lt, 162(, 163(
types of, 164
Bronze flake, in pigment, particle size
of, 180
Bronze pigment, as generic name for
metallic flake pigment, 180
Bronze powders, 179-80
Bronzing
color of
incident angle and, 164
measurement of, 164-68,
165(-167(
frequency of occurrence of, 168
types of
interface, 164
interference, 164, 168
Burnt umber
absorption and scattering of, 72
added to blue, reflectance curves for,
72-73, 270(, 272(
added to green, reflectance curves for,
75,285(
added to orange, reflectance curves
for, 291(
added to purple, reflectance curves
for, 75
added to red, reflectance curves for, 74,
275(, 277(, 279(
added to violet, reflectance curves for,
288(
added to yellow, reflectance curves for,
74, 281(, 283(
mixtures with chromatic colorants,
chromaticity diagram for, 268(
reflectance curves for, 72, 228,
266(, 267(
with Ti0
2
, reflectance curves for, 72
CIE notations for, 263t
Munsell notation for, 263t
Cadmium orange, reflectance curve for,
233-34,235(
Cadmium pigments, reflectance curves
for, 229
Cadmium red light, reflectance curve for,
227, 228(, 229
Cadmium red medium
as glaze, reflectance curves for, 108,
110(, 131
reflectance curve for, 227, 228(, 229
Cadmium yellow light, reflectance curve
for, 230(
Calibration, of spectrophotometer, 12
Carbazole dioxazine violet
chromaticity diagram of, 262(
in mixtures with Ti0
2
CIE notations for, 287t
Munsell notations for, 287t
reflectance curves for, 234(, 287(,
288(, 289(
at various concentrations, plot-
ted in various ways, 148(-154(
sensitivity to peroxides, 236
Carbon black
bronzed sample of
chromaticity coordinates of,
166, 167(
specular reflectance measurements
of, 164-67, 167t
in mixtures with Ti0
2
reflectance curves for, 264(, 265(
Carmine
as glaze, absorption K/S curve for,
112, 113(
reflectance curve for, 227-28, 228(
Cerulean blue, reflectance curve for,
225,226(
Chroma, 13. See also High-chroma colors;
Low-chroma colors
designations for, 255
Munsell, 49
Chroma, Munsell, of bronzes, 162
Chromaticity coordinates, 10, 25-27, 27(
averages of, in data analysis, 238,
241-42,242(
Chromaticity diagram, 26, 27(, 28(
examples using, 29-33
importance of, in data analysis, 238,
241-42,242(
of masstone whites and white-black
mixtures, 263(
for subtractive colorant mixtures,
31-32,32(
of various colorants, 262(
Chrome green, 231
reflectance curve for, 231, 232(
Chrome oxide green, reflectance curve for,
231,232(
Chrome yellow
in chrome green, 231
darkening by, 94
as glaze, reflectance curves for,
107, 108(
mixed with black, tinting strength of,
130- 31
reflectance curves for, 107,
108(, 131
Chrome yellow light, reflectance curve for,
230(, 231
Chrome yellow medium, reflectance curve
for, 230(, 231
Chromophtal Red 3B.
See Anthraquinoid red
CIE notation
for blues, 268t
for green and green mixtures, 284t
for neutral mixtures, 293t
for orange and orange mixtures, 290t
for reds, 273t
for violet and purple mixtures, 287t
for whites, blacks, and burnt umber,
263t
for yellows, 280t
CIE system, 20-33
color-difference equations in, 33-38
light source in, 21-22, 22(
measurements in, consistency in, 45--46
and Munsell system, 51, 52-53, 52(
standard observer in
color response of, 23-26, 23(
distance of, 21
CIELAB equation, 34-36
CIELUV equation, 34
Cinnabar. See Vermilion
CMC equation, 38
Coatings Encyclopedic Dictionary (FSCT
1995), 221-22
Cobalt blue, reflectance curve for,
226,226(
Cobalt violet, 233
Cobalt yellow (Aureolin), reflectance
curve for, 229-30
Color
description of
in examination results, suggested
protocol for, 253t, 255
industry scales and methods, 154-55
single-number color scales, 157-58
paper industry methods, 155-57
three-dimensional, 14, 24
interdisciplinary nature of, 258
nonspecrral, on chromaticity
diagram, 27
nonuniform, measurement considera-
tions, 243, 249
perception of, 1-2, 16, 258
relationships of, psychological effects
in, 33-34
spectrophotometric curves and, 13-14
vs. colorant, 15-16
Color: Universal Language and
Dictionary of Names
(Kelly and Judd 1976), 53-56
Color balance, distortion in, vehicle yel-
lowing and, 102-6
diagnostic tool for, 105-6
Color change
exposure-related
colorant change and, 93-102,
CPl-3
vehicle or substrate change and,
102-6
Index
measurement of, instrumentation
for, 247
Color control, inadequate, paint mixture
preparation and, 145--46
Color difference
acceptability considerations, 39
equations for, 34-38
exposure-related, 39--40
of identical materials, instrumentation
for, 247
perception of, 33-34
tristimulus filter colorimeters for,
38-39
Color-difference meters, 38-39. See also
Tristimulus filter colorimeters
The Color Harmony Manual, 56-57
Color matches. See Kubeika-Munk equa-
tions; Metamerism
Color-matching computer programs
for studies of pigment change, 93-94
for studies of vehicle yellowing,
104-5
using single constant (K/S), 85-86
using two constants (K and S), 69,
89, 92
Color-matching functions, 23, 23(
"Color Measuring Instruments: a Guide to
Their Selection" (1971), 245
Color names, 53- 56, 54(
generic, Colour Index, 296-97
Color notation systems
CIE. See CIE system
Munsell system, 49-53, 50(
OSA Uniform Color Scales, 56
Otswald, 56-57
universal color names, 53-56
Color science, instruction in, 258
Color Standards and Color Nomenclature
(Ridgway 1912), 55
Colorant(s)
absorption and scattering properties,
59-61
achromatic. See also Blacks; Whites
absorption and scattering properties
of, 70-72
changes in
Kubeika-Munk formulas for,
93-102
complete hiding, importance of,
95-96
log absorption K/S curves in,
98-100, 100(
measurements of, 100-102
perceived, 96-98
vehicle or substrate, 102-6
small, over time, 101-2
concentrations of
calculation of. See Kubeika-Munk
equations
relative, in mixtures, 81-86
heavy metals in, 135
small amounts of, detection and identi-
fication of, 66-67
tinting strength of, 128--44
vs. color, 15-16
353
Colorant composition
and color names, discrepancies in, 299
measurement of, instrumentation
for, 247
Colorant Mixture Computer
(COMIC), 85
Colorcurve system, 57
Colorimeters, features of, summaries of,
245--46. See also Tristimulus
filter colorimeters
Colorimetric metamerism, 48
Colorimetry, 15
additive color mixtures vs. subtractive
colorant mixtures in, 15-18,
16(, 19-20
definition of, 1-2, 15
lack of absolute measurements in, 12
primary colors in, 18-20, 19(
systems of, 15
uses of, 2
Colors, high-chroma. See High-chroma
colors
Colour Index generic names, 296-97
Commission Internationale de l'Eclairage,
2. See also CIE entries
Computers. See also Color-matching
computer programs
machine color vision in, 250
Consolidating resin, and surface reflection,
195-96, 196(
Continuous (analog) curve, from
spectrophotometer, 6-7
Contrast gloss, 203--4
Contrast ratio
definition of, 115
and hiding, 126-28, 127(
Copper, polished, protective coating for,
162-64
Craker, W. E., on scattering, 119
Critical angle, in paint-to-air interface, 88
Critical pigment volume concentration
(CPVC), 69
Cromophtal Scarlet Red. See Disazo
condensation
Dalamar Yellow. See Pigment Yellow 74
Damar varnish
and fluorescence, with age, 218
yellowing of, 104, 105(
Dark colors
measurement considerations, 23 7
surface reflection and, 191
Darkening, 94. See also Colorants,
changes in
from insufficient wetting, 98
log absorption curves and,
98-100, 100(
of vehicle or substrate, 102
Data. See also Measured data;
Spectrophotometric data
measured, analysis of
importance of, 244, 257
special considerations in, 238--43
precision of, 245
Densitometer, 151
354
Detectors. See Photodetectors
Dianisidine red
absorption band number and wave-
length of, 300t
reflectance curves for, 3 l 8f
Dictionary of Color (Maerz and Paul
1930), 55
Dielectric materials (nonmetals)
measurement of, instrument selection
for, 246-47
paint-to-air interface for, 86
surface reflection of, 8
Differential spectral curve analysis, 85, 105
Diffuse reflectance, 5
curve shapes, by color concentration,
147-49, 148f
measurement of
with bidirectional geometry, 8
instrument selection for, 249
of metals, in color measurements, 160
DIN (Deutsche Industrie Norm)
system, 57
Disazo condensation pigment
absorption band number and wave-
length of, 300t
reflectance curve for, 315 f
Dispersion, of pigments, 58
Distinctness-of-image gloss, 203-4
Dominant wavelength (A.d), in CIE system,
28-29, 29t
Dyes and dye solutions, 58
collections of, record of, 259
opacity of, 58
small concentrations of, detection and
identification of, 66-67
tinting strength of, 128-29
Emerald Green, darkening of, 94, 98-99,
99f, lOOf
color-matching for, 105
Excitation purity (pe), in CIE system,
28-29, 29t
Exposure
bronzing from, 168
and color change, 39-40
colorant change and, 93-102,
CPl-3
vehicle or substrate change and,
102-6
measurement considerations, 240
and surface reflection, 205
and vehicle, 102
yellowing, 102-6, l03f-l06f
Extender pigments, 221-24
common, 222t
Fade-Ometer
alizarin lake exposed in, 94, 95f,
CPl-2
Emerald Green exposed in, 98, 99f
Green Earth exposed in, 98, 99f
vermilion exposed in, 94, 96f
Fading, 94. See also Colorants, changes in
perceived, 96-98
Index
Feller, R. L., on darkening, 94, 98, 100
Filler, 222
Film sensitivity, metamerism and, 41
Filters
broadband, for tinting strength mea-
surements, 140-41
millipore, as working standard, 11
neutral-density
for fluorescent sample measure-
ments, 207
for interference pigments, 10
for metallic flake pigments, 183-85
transmittance calculations for, 63
ultraviolet, and fluorescent colorants,
213-14, 219
Fish-scale nacreous material, 172
Flavanthrone Yellow
as glaze
absorption of, 109-10, 131
reflectance curves for, 107-8, l09f
scattering of, 131
reflectance curve for, 229, 230f
Fluorescence, 205
in aging resins, 218-19
color measurement of, 206-7
measurement of, instrument selection
in, 248
and spectrophotometer design, 7
Fluorescent colorants
as additive mixtures, 205-6, 206f
ultraviolet filters and, 213-14, 219
Fluorescent pigments
historical use of, 207-9
modern, high-visibility, 210-14,
212f-217f, CPS
history of, 210-11
spectral curves of, 210-14,
212f-217f
CIE notation for, 212, 2l3t
substrate for, 215. See also Fluo-
rescent whitening agents (FWAs)
uses of, 214-17
substrate for, 215
ultraviolet illumination for, 215-17
Fluorescent whitening agents (FWA),
217-18
Forward scattering, 118
Frequency, in plotting graphs, 5
Fresnel reflection, 8. See also Specular
reflection
at interface
for Kubeika-Munk calculations, 79
in transparent materials, 63-64
for scattering by microvoids, 219-20
Fresnel's law, 63, 64
Fusion point, 20
FWA. See Fluorescent whitening agents
Gamboge, 230
in Hooker's Green, 232
General Electric Recording
Spectrophotometer, 12
Geometric metamerism, 41, 47-48
of matte surfaces, 192-96
Glazes
critical concentration of, for maximum
chroma, 113-14, CP3
Kubeika-Munk equations for,
106-14, 107
optical effects of, 109, A19, A21
over white, reflectance curves for,
107-9, l08f-lllf
type III colorant as, color of, 131
yellowed varnishes as, 110-12,
l l2f, l l3f
Gloss
and color change, 96-97, 101
color of
of bronzed rutile paint films, 168
vs. other surface types, 196-203,
l97f-202f, CP4
exposure and, 205
measurement of, 203-4
preferences for, 204
surface reflection of, and color, 191
types of, 203-4
Gloss white, as extender pigment, 222
Glossmeter, 203-4
and surface reflection, changes in, 101
Gold, hue of, 164
Gold leaf, 179
substitutes for, 179-80
Gold paste, reflectance curve for, 230f
Gonio mode, 9
Goniophotometer, 9
Goniospectrophotometer, 9-11. See also
Trilac goniospectrophotometer
for bronze, 164-68, l65f
features of, summaries of, 245-46
in metameric pair measurements,
47-48
Graphs
frequency and wavelength in, 5
importance of plotting measured data
on,238,240,242
Grayness
from additive mixtures, 17, l 7f
from complementary colors,
18-19, l9f
Green(s)
CIE notation for, 284t
Munsell notation for, 284t
reflectance curves for, 7 4-7 5,
285f-286f
Green Earth
darkening of, 98-99, 99f, lOOf
color-matching for, 105
reflectance curve for, 232f, 233
Green Gold (Nickel Azo Yellow),
reflectance curve for, 232f
Halogenated Anthanthrone, reflectance
curve for, 233, 235f
Handbook of Colorimetry (Hardy), chro-
maticity diagrams in, 30
Hansa Yellow G, reflectance curve
for, 230f
Haze, 124
Hiding
incomplete. See also Translucency
definitions related to, 114-15
in glazes, 106-14, CP3
Kubeika-Munk equation for,
107, 117
opacity calculations for, 116-18
scattering and, 118-24
microvoids and, 97, 97f, 219-21
reflectance and, 67
in color change studies, 95-96
spectrophotometric, 126-28, 127f
Hiding power, 125-28
definition of, 115
High-chroma colors
measurement considerations, 237
nonmetameric character of, 46
surface reflection and, 191, 196-203,
197f-202f, CP4
Hooker's Green, reflectance curves
for, 232
Hue
on chromaticity diagram, 26-27, 27f
designations for, 255
differences in, importance of, in color-
difference specifications, 39
reflectance maximum and, 14
Hue, Munsell, 49-50
of bronzes, 162, 163f
Hunter Color Difference Meter, 39
Hydrated chrome oxide, reflectance curve
for, 231, 232f
Identification
of object, record of, suggested protocol
for, 252-54, 253t
of pigment, 66-67, 70-78,
143-44, 144f
record of, suggested protocol for,
253t, 254- 55
Illuminant(s), and color perception, 1-2.
See also Light source
Illuminant A, 21
Illuminant B, 21
Illuminant C, 21
Illuminant D65, 21
Illuminant metamerism, 40-41, 42-47
Illumination, angle of
and metamerism, 41, 47-48
notation for, 9-10
Incident angle (i), 10
Indanthrone blue
reflectance curve for, 14 lf,
226f, 227
tinting strength of, 141-43,
141f, 142t
Indian red oxide, reflectance curve for,
228,228f
Indian Yellow, 208-9, 209f-21 lf
reflectance curve for, 229-30
Indigo
composition of, analysis of, 143-44,
144f, 298-99
reflectance curve for, 226f, 22 7
Index
Indofast Brilliant Scarlet
absorption band number and wave-
length of, 300t
perylene vermilion, reflectance curves
for, 327f
reflectance curve for, 323f
Industry scales and methods, 154-55
Inert pigment, 222-24
Infrared region, pigment reflectance in,
255-56
Inorganic pigments
absorption and scattering by, particle
size in, 133-35, 134f
as inert pigments, 222-24
Instruments, color-measuring. See also
Colorimeter; Spectrophotometer
cost of, 250
precision of, 245
in reports, 251
selection of, 244, 245-46
sample characteristics and, 246-50
Integrating sphere geometry, in spec-
trophotometers, 8
to measure perceived color change,
100-101
Integrating sphere spectrophotometer
for bronze measurements, 164-68,
166f, 167f
for fluorescent sample measurements,
vs. bidirectional geometry, 207
and metameric matte surfaces, 193-94
and translucency measurements,
121-22, 12lf
Intensity, of light beam, measurement of,
5,6
Inter-Society Color Council (ISCC), 53
Interface bronze, 164
Interfaces
air-sample, and reflectance, 7, 79
paint-to-air, and reflectance, 86-88,
87f, 88f
pigment-vehicle, changes in, perceived
color change from, 97-98, 97f
in transparent materials, specular
reflection in, 63-64
Interference bronze, 164
color of, 164, 168
Interference color. See Iridescence
Interference pigments, neutral-density
filter for, 10
Internal reflectance, in Kubeika-Munk
equation, 86
Internal surface reflection, 86-88,
87f, 88f
International Commission on Illumination
(CIE), 2. See also CIE entries
International Lighting Vocabulary (CIE
1987), 21
Irgazin, 229
Iridescence
definition of, 169
illumination of exhibits and, 179
measurement of, instrument selection
for, 248
in nature, 169, 173-74, 179
structural ca uses of, 169-71,
l 70f-171f
structural deterioration of, 179
Iridescent pigments
color of, 173, l 74f
355
over highly reflective substrate, 174
combined with other pigments,
177- 78
variable angle geometry in color
measurement of, 9
with transparent absorbing pigment,
174-77, l 77f, l 78f
Iron blue (Prussian blue)
bronzing effect by, 168
in chrome green, 231
in Hooker's Green, 232
reflectance curve for, 225, 226f
tinting strength of, 13 8
ISCC-NBS Method of Designating Colors
and a Dictionary of Color
Names (Kelly and Judd
1955), 53
ISCC-NBS Universal Color Language,
53-56
Isoindolinone-cobalt complex, 229
Isotropic scattering, 118
Isoviolanthrone violet, reflectance curve
for, 234f
Japanese prints, fugitive colorants in,
changes in, measurement
of, 101-2
Kubeika-Munk equations, 67-70
applications of
in calculation of tinting strength,
128-44
in cases of incomplete hiding,
glazes, 106-28
in paper industry, 151-54,
156-57
in study of colorant changes,
93-102
in study of vehicle yellowing, 102-6
extender pigments treatment in, 224
hyperbolic mathematical solutions
to, 125
qualitative application of, 70-78,
147-49
quantitative application of
complex estimation in (two-
constant), 86- 92
simple estimation in (single-
constant}, 78-86
quantitative limitations on, 145-46
Kumar, R., on solution spectrophoto-
metric technique, 66- 67
Lakes, 59
extender pigments in, 222
organic chemicals on inorganic
carriers, color of, 137, 137f
Lambert's law, 62-64
356
Laminar structures, iridescence in,
170-71, 170(, 171(
Lead, in colorants, 135
Lead antimonate, reflectance curve for,
229-30
Lead carbonate, nacreous, 172
Lead-tin yellow, reflectance curve for,
229-30
Leres, Trilac goniospectrophotometer by,
9, 10(
Light, interference of. See Iridescence
Light aluminum hydrate, as extender
pigment, 222, 223
Light source
auxiliary, filtered, for color balance
distortions, 105-6
for fluorescent sample measurements,
206-7,212,212(-217(
instrument selection for, 249
intensity of, measurement, 5-6
for spectrophotometers, 6
location of, 7
Lighting, for fluorescent pigments,
215-17, 248, CPS
Lightness. See also Y (Tristimulus value)
in CIE system, 28, 29t
designations for, 255
differences in, importance of, in color-
difference specifications, 39
reflectance and, 14
Lignin, yellowing curve of, 104, 106(
Linseed oil, and fluorescence, with age,
218-19
Lithol maroon, absorption band number
and wavelength of, 300t
Lithol Red
absorption band number and wave-
length of, 300t
diffuse reflectance curve for, 307(
Lithol Rubine, reflectance curve for,
227,228(
Log absorption curves. See Absorption
K/S curves, on logarithmic scale
Low-chroma colors
measurement considerations, 240
metamerism and, 46
Luminescence, 205
Luster. See Pearlescence
Machine color vision, 250
Malachite, 233
Manganese blue, reflectance curve for, 227
Manganese violet, 233
Masstone(s), spectral reflectance of,
69-70, 73
in pigment calibration (for K and S),
89-91
Mastic, and fluorescence, with age, 218
Matte Paint (Hansen, Walston, and
Bishop), 191
Matte surfaces
color of, vs. other surface types,
196-203, 197(-202(, CP4
exposure and, 205
Index
geometric metamerism and,
192-96
sphere-mode reflectance measure-
ments of, 193-94, l94t
variable-angle reflectance measure-
ments of, 192-95, 192(, 193(,
194t, 195(
measurement of, instrument selection
for, 249
microstructure of, 191
reflection from, 8
Maxwell disk, 17-18
Measured data
precision of, 245
reproducibility of, 12-13, 101-2, 245
suggested protocol for, 252-56, 253t
Measured-data analysis
averaging in, 240-43
curved surfaces, 243
importance of, 244, 257
mottled or uneven areas, 243
sampling problem, 240-43,
241(, 242(
principles in, 238
sample problem, 238-40, 239(, 239t
Medium. See Vehicle
Metal alloys, color of, influences on,
160-61, 161t, 164
Metal(s) (nondielectrics)
color of
evaluation of, 164
measurement of, 8, 158-59
heavy, in colorants, 135
reflectance of, 128
equation for, 158
measurement of, instrument selec-
tion in, 247-48
surface reflection of, 8
Metallic flake pigment
CIE notations for, 183t, 184
color of, 180
concentration of, 181
frequency of usage of, 180, 189
history of, 179-80
measurements of
instrument selection for, 248
variable angle geometry in, 9
orientation of flakes in, 181, 181(
particle size of flakes in, 180
reflectance of, 181-89, 181(-188(,
l83t
with scattering pigment, 186-89,
186(-188(
with transparent absorbing
pigment, 184(, 185-86, 185(
Metamerism, 40-49
geometric, 41, 4 7-49
of matte surfaces, 192-96
illuminant, 40-41, 42-47
indexes for, 42-43
and light source colors, 48
and low-chroma colors, 46
measurement of, 45-46
instrumentation for, 247
multiplicative corrections in, 44-45
observer, 40-41, 45
other names for, 48
and spectral curves, characteristics
of, 47
Microspectrophotometry, thin-layer
chromatography with, 67
Microvoids, 97, 97(, 219-21
uses for, 220-21
Mie theory, of scattering, 118, 128
and particle size, 134-38, 134(,
136(, 138(
Millimicron (m), 5
Millipore filters, as working standard, 11
Mirror-like reflection. See Specular
reflection
Molybdate orange
absorption band number and
wavelength of, 300t
chromaticity diagram of, 262(
CIE notation for, 290t
darkening by, 94
Munsell notation for, 290t
reflectance curves for, 76, 290(-292(,
291(, 292(, 304(
Monochromatic light, for fluorescent sam-
ple measurements, 206- 7, 212,
212(-217(
Monochromator, in spectrophotometers, 6
location of, 7
Monterey Red. See Red 2B-calcium
Mottled colors, measurement considera-
tions, 243
Multiple scattering, 59, 118
Munsell Book of Color, 49, 50(, 51-52,
52,53,255
Munsell Chroma, 49
of bronzes, 162
Munsell Hue, 49- 50
of bronzes, 162, 163(
Munsell notation
for mixtures with black, 263t, 268t,
273t, 280t; see also tables of
pigments listed below
BON red, 273t
in neutral mixtures, 293t
in orange mixtures, 290t
in purple mixtures, 287t
burnt umber, 263t
carbazole dioxazine violet, 287t
green mixtures, 284t
molybdate orange, 290t
neutrals, 293t
orange mixtures, 290t
Permanent Yellow FGL, 280t
in orange mixtures, 290t
phthalo green, 284t
in neutral mixtures, 293t
phthalocyanine blue, 268t
in neutral mixtures, 293t
purple mixtures, 287t
quinacridone red, 273t
in orange mixtures, 290t
in purple mixtures, 287t
red iron oxide, 273t
in neutral mixtures, 293t
in orange mixtures, 290t
in purple mixtures, 287t
ultramarine blue, 268t
in neutral mixtures, 293t
in purple mixtures, 287t
whites, 263t
yellow iron oxide, 280t
in neutral mixtures, 293t
in orange mixtures, 290t
Munsell system, 49-53, 50(
and CIE system, 51, 52-53, 52(
and universal color names, 53-54
Munsell Value, 49
of bronzes, 162
Nacreous lead carbonate, 172
Nacreous pigments. See Pearlescent
pigments
Nanometer (nm), 5
Naphthol red
absorption band number and wave-
length of, 300t
reflectance curve for, 309(, 312, 314(
National Bureau of Standards (NBS), 53.
See also National Institute for
Standards and Technology
National Bureau of Standards (NBS) unit,
36-37, 39
National Gallery, London, color change
monitoring at, 102
National Institute for Standards and
Technology, 53
Natural Color System (Sweden), 57
Natural madder lake, 207, 208(
NBS unit, 36-37, 39
Neutral-density filter
for fluorescent sample measurements,
207
for metallic flake pigments, 183-85
for pearlescent and interference
pigments, 10
Neutrals and near neutrals
from additive mixtures, 17, 17(
CIE and Munsell notation for, 293t
from complementary colors,
18-19, 19(
metamerism and, 46--4 7
reflectance curves for, 13, 77, 293(
Nickel Azo Yellow, reflectance curve
for, 232(
Nitrocellulose, 172
Nondielectrics. See Metal(s)
Nonmetals. See Dielectric materials
Object identification, protocol for,
252-54,253t
Observer metamerism, 40--41, 45
Opacity
calculation of, 116-18
definition of, 115
materials requiring knowledge
of, 128
Index
measurement of, 121-24
charts for, 124-25
relative, and particle size, 136, 136(
Opaque, definition of, 115
Opaque materials
measurement of, instrument selection
for, 246--4 7
reflectance of, 67-69
Optical density, 65, 149-54
measurement of, 66
Optical Society of America (OSA), 56
Orange(s)
CIE and Munsell notation for, 290t
hue of, wavelength characteristics
of, 231
reflectance curves for, 76, 233-34,
235(, 291(-292(
Organic materials, absorption by, 66-67
Organic pigments
inert pigments in, 222-24
scattering and absorption by, 59
particle size and, 135-37, 136(
small amounts of, solution spectropho-
tometry for, 143--44, 144(
Orono Red. See Red 2B-barium
Orthochlorparanitraniline
absorption band number and wave-
length of, 300t
reflectance curve for, 306(
Otswald color system, 56-57
Paint mixture preparation, and inadequate
color control, 145--46
Paint-to-air interface. See also Air-sample
interface
and reflectance, 86-88, 87(, 88(
The Painter's Methods and Materials
(Laurie 1926), 102
Pantone, Incorporated, color collection
of, 57
Paper
absorption and scattering considera-
tions, 151-52
color description methods for, 155-57
Paranitraniline red
absorption band number and wave-
length of, 300t
reflectance curves for, 319(
Pearlescence
definition of, 169
measurement of, instrument selection
for, 248
Pearlescent pigments
combined with other pigments,
174-78, l 77(, l 78(
history of, 172-73
neutral-density filter for, 10
Percent reflectance, 5
Perceptual systems, 15, 20
Perfect reflecting diffuser, 11
Permanent Red 2B. See Red 2B-barium
Permanent Yellow FGL
chromaticity diagram of, 262(
CIE notation for, 280t
in green mixtures
CIE notation for, 284t
reflectance curves for, 286(
Munsell notation for, 280t
in orange mixtures
357
CIE and Munsell notation for, 290t
reflectance curves for, 292(
reflectance curves for, 74, 282(, 283(
Peroxides, sensitivity of carbazole
dioxazine violet to, 236
Perylene maroon
absorption band number and
wavelength of, 300t
reflectance curve for, 322(
Perylene red, reflectance curves for, 228(,
229, 322(
Perylene vermilion
absorption band number and
wavelength of, 300t
reflectance curves for, 322(, 32 7(
Phloxine
absorption band number and
wavelength of, 300t, 3 l3t
reflectance curve for, 313(
Phosphorescence, 205
Photodetectors, in spectrophotometers, 6
Photographs, metamerism and, 41
Phthalocyanine blue
and benzidine yellow, interaction
between, 234-35, 236(
bronzed sample of
chromaticity coordinates of,
166, 167(
specular reflectance measurements
of, 164-67, 165(, l67t
chromaticity diagram of, 262(
CIE notation for, 268t
in green mixtures, reflectance curves
for, 286(
Munsell notation for, 268t
in neutral mixtures
CIE and Munsell notation
for, 293t
reflectance curves for, 293(
reflectance curves for, 72-73, 225,
226(, 271(, 272(
in violet mixtures, reflectance curves
for, 289(
Phthalocyanine green
bronzed sample of
chromaticity coordinates of,
166, 167(
specular reflectance measurements
of, 164-67, l67t
chromaticity diagram of, 262(
CIE notation for, 284t
Munsell notation for, 284t
in neutral mixtures
CIE and Munsell notation
for, 293t
reflectance curves for, 293(
reflectance curves for, 74-75, 232,
232(, 284(-285(, 285(, 286(
Physical systems, 15
358
Pigment(s). See also Type I; Type II;
Type III
absorption by, 14, 58-61
changes in, Kubeika-Munk formulas
for, 93-102
chromaticity diagrams for, 31-32, 32f
collections of, record of, 259
concentration of
log absorption curves and, 98, lOOf
reflectance and, 67-70, 80-86
and fading or darkening, 93-96,
CP2
at incomplete hiding, 118
and scattering, 118-20
small, detection and identification
of, 66-67, 143--44
interactions among, 234-36
maximum Chroma of, 69, 113-14,
CP3
metamerism and, 41--4 2
names for, discrepancies in, and
composition, 299
newer, focus on, 257
particle size of, 60, 6lf
metallic flakes in, 180
tinting strength of, 133-39
prices of, tinting strength comparisons
and, 141--43, 14lf, 142t
primary, reflectance curves for, 72-7 4
scattering by, 58-61
secondary, reflectance curves for,
74-76
tinting strength of, 129--44
in mixtures, 130- 33
particle size in, 133-39
Pigment Green B (Pigment Green 8),
reflectance curves for, 232, 232f
Pigment Handbook (Patton 1973), 60
on gold leaf substitutes, 179
on pearlescent and iridescent pigments,
172
Pigment mixtures
absorption and scattering of, 69,
73, 74
tinting strength and, 130-33, 132f
reflectance of, 69, 73, 74
relative concentrations of, 81- 86
units used in, 84
Pigment-vehicle interface, changes in, per-
ceived color change from,
97-98, 97f
Pigment volume concentration (PVC),
low, 98
Pigment Yellow 74 (Dalamar Yellow),
particle size vs. tinting strength,
135- 36, 136t
Plotting graphs
frequency in, 5
importance of plotting measured data
on,238,240,242
wavelength in, 5
Pointillists
additive mixtures from, 19-20
measurement problems with, 243
Index
Polarization, and specular reflection
calculations, 64
Polychromatic light, for fluorescent sam-
ple measurements, 206- 7, 212,
212f- 217f
Polymer matrix, air spaces in. See
Microvoids
Pressed barium sulfate (BaS0
4
), 10,
11, 12
in matte surfaces, surface reflectance
of, 194
Pressed poly(tetrafluoroetylene)
(Halon G-50), 11
Primary colorants
reflectance curves for, 72-74, 225-31,
226f-230f, 268f- 283f
vs. primary colors, 15, 18, 20f
Primary colors
complementary, 18-19, 19f
vs. primary colorants, 15, 18, 20f
Primrose chrome yellow, reflectance curve
for, 230f, 231
Principles of Color Technology (Billmeyer
and Saltzman 1981), 52, 329
Prussian blue. See Iron blue
Psychophysical systems, 15, 20. See also
CIE system
Purity
estimation of, 27-28, 29t
examples of, 29t
Purple(s)
CIE notation for, 287t
Munsell notation for, 287t
reflectance curves for, 75- 76, 233,
234f, 287f-289f
Purpurin, in natural madder lake, 207
PVC. See Pigment volume concentration
Pyranthrone, reflectance curve for,
228f, 229
Pyrazolone red
absorption band number and wave-
length of, 300t
reflectance curve for, 228f, 303f, 315f
Quinacridone red
absorption band number and wave-
length of, 300t
chromaticity diagram of, 262f
CIE notation for, 273t
as glaze
absorption of, 109-10
reflectance curves for, 108,
lllf, 131
Munsell notation for, 273t
in orange mixtures
CIE and Munsell notation for, 290t
reflectance curves for, 292f
in purple mixtures
CIE and Munsell notation for, 287t
reflectance curves for, 289f
reflectance curves for, 73-74, 228f,
278f, 279f, 326f
Quinacridone violet, reflectance curve
for, 234f
Radiance, and fluorescence, 205-6
Radiometry, 250
Red pigments
absorption band number and wave-
length for, 300t
hue of, wavelength characteristics
of, 231
reflectance curves for, 72, 73- 74,
227-29, 228f, 302f- 327f
Red 2B-barium
absorption band number and wave-
length of, 300t
reflectance curve for, 228f, 302f,
303f, 317f
Red 2B-calcium
absorption band number and wave-
length of, 300t
reflectance curve for, 31 Of
Red 2B-strontium
absorption band number and wave-
length of, 300t
reflectance curve for, 31 lf
Red iron oxide
chromaticity diagram of, 262f
CIE notation for, 273t
Munsell notation for, 273t
in neutral mixtures
CIE and Munsell notation for, 293t
reflectance curves for, 293f
in orange mixtures
CIE and Munsell notation for, 290t
reflectance curves for, 292f
in purple mixtures
CIE and Munsell notation
for, 287t
reflectance curves for, 289f
reflectance curves for, 73, 274f, 275f
characteristics of, 227, 228, 228f
Reference standards, 10, 11-12, 256
condition of, importance of, 13
Reflectance
definition of, 5, 115
diffuse. See Diffuse reflectance
at incomplete hiding. See Hiding,
incomplete
Kubeika-Munk equation for, 67-70,
86- 92
simplified, 80-84
measurements of, reproducibility con-
siderations, 12-13, 101-2
pigment concentration and
iterative method for, 84-85
Kubeika-Munk equation for, 67-70,
80-84
with colorant change, 93-102
total. See Total reflectance
in ultraviolet and infrared regions,
255- 56
Reflectance curve(s)
library of, need for, 77
plotted with K/S on log scale, 147--49,
148f- 154f
vs. digital output, of spectrophotomet-
ric data, 6-7
vs. log absorption K/S curves, in expo-
sure-related color changes,
98-100, 100(
Reflectance curve analysis
computer aids for, 85
qualitative pigment identification with,
70-78
Reflectance factor, 5
definition of, 115
Reflection
definition of, 115
surface. See Surface reflection
surface type and, 7-8, 79-80
Reflection density, 149-54
Refractive index
and reflection at air-object interface, 79
and reflection at paint-air interface,
87-88, 87(, 88(
and scattering and absorption by pig-
ments, 58-61
spectrophotometer design and, 7-8
and surface reflection, 190-91
Relative reflectance, 10
of metallic paints, 183, 184
Repeatability, 245
Reports, 251
suggested protocol for, 252-56, 253t
Reproducibility, 12-13, 101-2, 245
Resin(s)
consolidating, and surface reflection,
195-96, 196(
and fluorescence with age, 218-19
Results, of examination, suggested proto-
col for, 252-56, 253t
Retroreflection, 250
Robinson, F.D., on scattering, 119
Rutile Ti0
2
. See Titanium white
Sample(s)
characteristics of, 7
in instrument selection, 246-50
condition of, importance of, 13
examination results, protocol for,
252-56,253t
identification of, protocol for, 252-54,
253t
location of, 7
size of, measurement considerations,
242,249
Samples, size of, analysis considerations,
66-67
Saturation
on chromaticity diagram, 26, 27(
differences in, importance of, in color-
difference specifications, 39
low, measurement considerations, 240
and spectral curves, 13-14
Saunderson, J.L., on paint-to-air interface,
86-87
Scattering, 118-24
as colorant characteristic, 58-61
in Kubeika-Munk equations,
78-79, 88
microvoids and, 97, 97(, 219-21
Index
Mie theory, 118, 128
multiple, 59, 118
surface. See Smoothness
and tinting strength calculations
particle size and, 133-39
for pigment mixtures, 130-33
Scattering-absorption relationship, and
reflectance, 67-70, 68(
Scattering coefficient (S), in Kubelka-
Munk equation, 88-92
SCE. See Specular component of the
reflectance excluded
SCI. See Specular component of the
reflectance included
Secondary additive colors, 18, 19(
Secondary colorants, reflectance curves
for, 74-76, 231-34, 232(-235(,
284(-293(
Sheen, 203--4
Sight
color perception and, 1-2
additive color mixtures and, 16
spectral sensitivity and, 22-23, 22(
Single-constant (K/S) Kubeika-Munk
calculations, 78-86
Single scattering, 118
Small samples
analysis of, techniques for, 66-67
measurement considerations,
242,249
Smalt, reflectance curve for, 22 7
Smoothness, of metals, 160, 162
Solution spectrophotometry, 65-67,
143--44, 144(
Solvents
dilution of textile dyes with, absorp-
tion curves and, 14
microvoid formation from, 221
Special Metamerism Index: Change of
Illuminants (CIE 1973), 42
Spectral curve. See also
Spectrophotometric curve
definition of, 1, 5
Spectral data, 244
suggested protocol for, 252-56, 253t
Spectral reflectance curve(s)
library of, need for, 77
plotted with K/S on log scale, 147--49,
148(-154(
vs. digital output, of spectrophotomet-
ric data, 6-7
vs. log absorption K/S curves, in expo-
sure-related color changes,
98-100, 100(
Spectral reflectance curve analysis
computer aids for, 85
qualitative pigment identification with,
70-78
Spectral response. See also Color-matching
functions
characteristics of, 22-23
in tristimulus values calculation, 24
Spectrocolorimeter, 139--40
Spectrogoniophotometer, 9
Spectrophotometer, 5, 6-11
abridged, 6
359
vs. tristimulus filter colorimeters, 39
calibration of, 12
continuous curves vs. digital output
of, 6-7
design of, 7-9
features of, summaries of, 245--46
for fluorescent samples, 206-7
to measure perceived color change,
100-101
measurements made by, 6-7
parts of, 6
reliability of measurements by, 12-13
in tinting strength measurement, 139
wavelength range of, 250
Spectrophotometric curve, 13-14. See also
Spectral curve
definition of, 1, 5
reference library of
need for, 258-59
suggested organizational scheme
for, 298-301
Spectrophotometric data, 244. See also
Measured data
reflectance curves vs. digital output,
6-7
suggested protocol for, 252-56, 253t
Spectrophotometric hiding, 126-28, 127(
Spectrophotometry, 244--45
definition of, 1
lack of absolute measurements in, 12
precision of, 245
solution, 65-66, 143--44, 144(
uses of, 2
Spectroradiometry, 249-50
Spectrum locus, in chromaticity diagram,
26,27(
Specular component of the reflectance
excluded (SCE), 160
and surface scattering, 161t, 162
Specular component of the reflectance
included (SCI), 160
Specular gloss, 203--4
Specular reflection, 8. See also Fresnel
reflection
of bronzed dialectric materials,
measurement of, 164-68,
165(-167(
at interface
for Kubeika-Munk calculations, 79
in transparent materials, 63-64
of metals, calculation of, 160
for nondielectric materials
calculation of, 160
equation for, 158
Stainless steel flake pigment, 180
particle size of, 180
Standard observer
in CIE system, 21, 23-26, 23(
and metamerism, 40
Standards. See Reference standards
Staniforth, S., on metameric matches,
41--42
360
Substrate
change in, exposure-related, 103-6,
103(-106(
at incomplete hiding. See Glazes
extender pigments in, 224
at incomplete hiding, 114-28. See also
Glazes
paper as, absorption and scattering
considerations, 151-54
Subtractive colorant mixtures, 15-17, 16(
chromaticity diagrams for, 31-32, 32(
primary colors for, 18, 19(
Surface
changes in, perceived color changes
from, 96-98
log absorption K/S curves and,
98-100, 100(
measurement of, 100-102
curved, measurement considerations,
243
textured, color of, vs. other surface
types, 196-203, 197f-202f, CP4
Surface differences, and high-chroma
color, 196-203, 197(-202(, CP4
Surface polish. See Smoothness
Surface reflection, 7-8, 79-80
and color change, 96-97, 101
consolidating resin and, 195-96, 196(
exposure and, 205
of glossy surfaces, measurement con-
siderations, 203-4
and high-chroma colors, 191,
196-203, 197(-202(, CP4
internal, 86-88, 87(, 88(
in Kubelka-Munk calculations, 79-80
macro, 201-3
measurement considerations, instru-
ment selection and, 24 7
micro, 191-96
from perpendicular incident light,
63-64, 79
refractive index differences and, 190-91
of transparent materials, and transmit-
tance, 63-64
Surface scattering. See Smoothness
Synthetic red oxide
absorption band number and wave-
length of, 300t
reflectance curve for, 228, 228(,
274(, 305(
Synthetic yellow iron oxide, 230, 230(
Technical Association of Pulp and Paper
Industries (TAPP!), 154
brightness standard, 155-56
Terminology, of appearance, 115
Textile dyes
absorption of, 14
opacity of, 58
Textile pigments, absorption and scatter-
ing by, 59
Textured surface, color of, vs. other
surface types, 196-203,
197(-202(, CP4
Index
Thin-layer chromatography, with
microspectrophotometry, 67
Thioindigo red
absorption band number and wave-
length of, 300t
reflectance curve for, 228(, 229,
234(, 320(
Tinting strength, 128-44
in comparing pigment prices, 128,
141-43, 141(, l42t
equations for, 129-30
measurement techniques, 139-42
particle size and, 134-39
of pigment mixes, 130-33
Titanium white (Ti0
2
)
anatase, absorption and scattering
properties of, 70, 71
in matte surfaces, surface reflectance
of, 194
on mica flakes, in interference pig-
ments and pearlescent flakes,
172-73
rutile
absorption and scattering properties
of, 70-72
bronzing in, 168
tinting strength of, 138-39, 138(
Toluidine red, reflectance curve for,
227,228(
Toluidine yellow, reflectance curve
for, 230(
Toners, 59, 137
Total reflectance, 5
of metals, in color measurements, 160
Trade organizations, color description
standards, 154-55
Transfer standard
definition of, 11
materials used for, 11
Translucency. See also Hiding, incomplete
materials requiring knowledge
of, 128
measurement of, 121-24
instrument selection for, 246
of thick materials, 123
of thin materials, 123-24
scattering and, 118-20
and visual appearance difference,
115-16
Translucent, definition of, 115
Transmission, definition of, 115
Transmit, definition of, 115
Transmittance
definition of, 115
measurement considerations and, 66
instrument selection, 246
in translucent materials, 121-24, 121(
in transparent materials
absorption and, 64-67
thickness and, 62-64
Transparency, definition of, 115
Transparent, definition of, 115
Transparent liquids, single-number color
scales for, 158
Transparent materials
Beer-Bouguer equation for, 62,
64-67, 150
Bouguer's equation for, 62-64
instrument selection for, 246
interfaces in, specular reflection in,
63-64
iridescence in, 170- 71, 170(, 171(,
173, l 75(, l 76(
Trilac goniospectrophotometer, 9,
10(, 12, 246. See also
Goniospectrophotometer
to measure specular reflectance of
bronzes, 164-68, 165(- 167(
in metameric pair measurements,
47-48
Tristimulus colorimetric measurements,
precision of, 245
Tristimulus fi lter colorimeters, 38-39
features of, 245-46
filters in, 63, 155
for metameric measurements, 46
in tinting strength measurement,
139-42
Tristimulus values
averages of, 23 8
calculation of, 24-25, 25(
Tungstate green toner, reflectance curve
for, 232f
Two-constant Kubelka-Munk equation,
86-92
Type I, II, and III pigments defined, 59
Type I pigments
absorption and scattering by, 59, 70
particle size and, 138-39, 138(
tinting strength of, 129-30
Type II pigments
absorption and scattering by, 59,
69-70
particle size and, 137
as glazes, chromaticity of, 113-14
mixed with white, chromaticity of,
69, 113
tinting strength of, 130
in mixtures, 131-32, 132(
Type III pigments
absorption and scattering by, 59, 70
particle size in, 133-35, 134(
as glazes, chromaticity of, 113-14
tinting strength of, 130
in mixtures, 130-33, 132(
Ultramarine blue
chromaticity diagram of, 262(
CIE notation for, 268t
in green mixtures, reflectance curves
for, 286(
Munsell notation for, 268t
in neutral mixtures
CIE and Munsell notation for, 293t
reflectance curves for, 293(
in purple mixtures
CIE and Munsell notation for, 287t
reflectance curves for, 289(
reflectance curves for, 72-73, 141(,
225, 226(, 227, 269(, 270(
tinting strength of, 141-43, 141(, 142t
Ultraviolet filters, and fluorescent
materials, measurement of,
213-14, 219
Ultraviolet illumination, and fluorescent
pigments, 215-17, 248
Ultraviolet region, pigment reflectance in,
255-56
Uniform materials, measurement of,
instrument selection for,
246-47
Universal Color Language, 53-56
Universal scattering curve, 138, 138(
Value, Munsell, 49
of bronzes, 162
Vandyke Brown, 299
Variable angle geometry
for metamerism of matte surfaces,
192-95, 192(, 193(, 194t, 195(
in spectrophotometers, 8, 9
Varnishes
and fluorescence, with age, 218-19
yellowed, 104, 105(, 106(
spectral absorption K/S curve analy-
sis of, 112, 113(
spectral reflectance curve analysis
of, 110-11, 112(
Vehicle
aromatic groups or solvents in, phthalo
blue contamination in, 235
to consolidate surface, and surface
reflection, 195-96, 196(
exposure-related change in, 102
yellowing, 102-6, 103(-106(
at incomplete hiding. See Glazes
extender pigments considered in, 224
and fluorescence, with age, 218
Verdigris, 233
Vermilion
darkening of, 94, 96(
reflectance curve for, 96(, 229
Vert Emeraude Dull. See Chrome oxide
green
Vert Emeraude Transparent. See Hydrated
chrome oxide
Vesiculated beads, 221
Viewing angle (a), 10, 10(
and metamerism, 41, 47-48
Index
Viewing distance, and additive mixtures,
19-20
Violet(s), 233, 234(. See also Carbazole
dioxazine violet; Purples
Violet-blue fluorescent dyes, as FWAs, 218
Viridian. See Hydrated chrome oxide
Vision
color response and, 22-23, 22(
as integrative process, 1, 16
Voids, microscopic. See Microvoids
Vulcan Fast Red. See Dianisidine red
Waco Red
absorption band number and wave-
length of, 300t
reflectance curve for, 312(
Wavelength
in phase/out of phase, in interference
color, 169, 170(
in plotting graphs, 5
and refractive index, 191
unit of, 5
Wavelength range, of instruments, 250
Wetting, lack of, color changes from, 98,
99-100
White(s)
CIE notation for, 263t
masstone
chromaticity diagram of, 263(
reflectance curves for, 266(
mixtures with blacks, chromaticity
diagram of, 263(
mixtures with chromatic colorants,
chromaticity diagram for, 268(
Munsell notation for, 263t
in pigment mixtures
absorption curves and, 14
and reflectance, 68, 263t, 268t,
273t, 280t, 284t, 287t, 290t,
293t
in single-constant Kubeika-Munk
equation, 78-79, 80
in two-constant Kubeika-Munk
equation, 88
reflectance curves for, 70-72,
89-91,266(
White standard. See also Reference
standards
types of, 11
Whiteness, indices for, 157-58
Whitening agents, fluorescent, 217-18
Working standards, 11-12
surface characteristics of, 11-12
X-ray diffraction, for white pigment
identification, 71
X-ray fluorescence, for white pigment
identification, 71
X (Tristimulus value), 24
Y (Tristimulus value), 24
Yellow
hue, wavelength characteristics
of, 231
reflectance curves for, 74, 229-31,
230(
scattering of, in mixtures, 88- 89
Yellow iron oxide
chromaticity diagram of, 262(
CIE notation for, 280t
in green mixtures
CIE notation for, 284t
reflectance curves for, 286(
Munsell notation for, 280t
in neutral mixtures
CIE and Munsell notation for,
293t
reflectance curves for, 293(
in orange mixtures
CIE and Munsell notation
for, 290t
reflectance curves for, 292(
spectral reflectance curves for,
280(, 281(
Yellow lead chromate. See Chrome
yellow
Yellowing
fluorescent whitening agents for,
206(, 218
of varnishes, 104, 105(, 106(
as glazes, 110-12, 112(, 113(
vehicle, 102-6, 103(-106(
compensation for, 105-6
Yellowness
indices for, 157-58
of paper, industry standards,
155-57
361
Z tristimulus filter, vs. special filter for
paper brightness measurements,
155-56
Z (Tristimulus value), 24
Illustration Credits
The following have generously granted permission to reprint previously published
material.
Figure 4.5, 4.24, and color plates 1, 2, and 3: From R. M. Johnston-Feller, "Reflections on
the phenomenon of fading,"Journal of Coatings Technology 58 (1986).
Figure 4.6: From R. L. Feller, "Studies of the darkening of vermilion by light," Report and
Studies on the History of Art (1967), National Gallery of Art, Washington, D.C.
Figures 4.8, 4.9, 6.12, and 6.13: From R. L. Feller, "Problems in reflectance spectropho-
tometry," IIC 1967 London Conference on Museum Climatology (1968), International
Institute for Conservation of Historic and Artistic Works, London.
Figures 4.10 and 4.11: From R. M. Johnston, "Spectrophotometry for the analysis and
description of color," Journal of Paint Technology 39 (1967).
Figures 4.14, 4.15, 4.16, and 4.17: From R. M Johnston and R. L. Feller, "Optics of paint
films: Glazes and chalking," Application of Science in Examination of Works of Art
(1967), Museum of Fine Arts, Boston.
Figures 4.20 and 4.21: From W. E. Craker and F. D. Robinson, "The effect of pigment
volume concentration and film thickness on the optical properties of surface coatings,"
Surface Coatings International 50 (1967).
Figure 4.26: From V. Buttignol, "Optical Behavior of Iron Oxide Pigments," Journal of
Coatings Technology 40 (1968).
Figures 5.5, 5.6, and 5.7: From R. M. Johnston and D. Osmer, "Exposure evaluation,
part II-Bronzing," Journal of Coatings Technology 51 (1979).
Figure 6.11: From F. Grum, "Colorimetry of fluorescent materials," Optical Radiation
Measurements, vol. 2, Color Measurement, ed. F. Grum and C. J. Bartleson; reprinted by
permission of Academic Press, New York.
Figures 7.1, 7.2, 7.3, 7.4, 7.5, and 7.6: From R. M. Johnston, "Spectrophotometry for the
analysis and description of color," Journal of Paint Technology 39 (1967).
Other acknowledgments will be found following figure legends.
About the Author
After graduating from the University of Illinois with a degree in chemistry,
Ruth Johnston-Feller spent the early years of her career in industry, con-
ducting research in infrared and ultraviolet spectrometry. With the advent
of improved measuring instruments and computing devices that made the
serious study of color practical in the late 1950s, she shifted her research
to the visible part of the spectrum and to the human response to color,
areas in which she had a long-standing interest. She joined the Pittsburgh
Plate Glass Company, and under her stewardship the company became
the first to use instrumental methods for formulating and controlling
color in paint and plastics manufacturing facilities. Later she served
as director of application services for the Color Systems Division of
Kollmorgen Corporation, creators of computer color-matching programs.
Her final industrial experience was with the Ciba-Geigy Corporation,
Pigments Division (now Ciba Specialty Chemicals, Colors Division).
While at the Pittsburgh Plate Glass Company, she met
Robert L. Feller, who introduced her to art conservation research. After
their marriage in 1975, she worked part-time with her husband's group
at the Research Center on the Materials of the Artist and Conservator
at the Carnegie Mellon Research Institute, Carnegie Mellon University.
There she devoted her attention principally to the measurement of
color and appearance, and to the study of fading of pigmented and
dyed materials.
Long active in many technical societies, Ruth Johnston-Feller
authored more than fifty technical publications. In 1985 she presented
the Mattiello Memorial Lecture of the Federation of Societies for
Coatings Technology, an honor given for her distinguished contributions
to the coatings industry. She received numerous awards in recognition of
her work, including the MacBeth Award, presented by the Inter-Society
Color Council for outstanding contributions to the science of color. She
died in 2000.
ISBN 0-8923b-58b-2